Artificial intelligence and machine learning‐aided drug discovery in central nervous system diseases: State‐of‐the‐arts and future directions

Abstract

Neurological disorders significantly outnumber diseases in other therapeutic areas. However, developing drugs for central nervous system (CNS) disorders remains the most challenging area in drug discovery, accompanied with the long timelines and high attrition rates. With the rapid growth of biomedical data enabled by advanced experimental technologies, artificial intelligence (AI) and machine learning (ML) have emerged as an indispensable tool to draw meaningful insights and improve decision making in drug discovery. Thanks to the advancements in AI and ML algorithms, now the AI/ML‐driven solutions have an unprecedented potential to accelerate the process of CNS drug discovery with better success rate. In this review, we comprehensively summarize AI/ML‐powered pharmaceutical discovery efforts and their implementations in the CNS area. After introducing the AI/ML models as well as the conceptualization and data preparation, we outline the applications of AI/ML technologies to several key procedures in drug discovery, including target identification, compound screening, hit/lead generation and optimization, drug response and synergy prediction, de novo drug design, and drug repurposing. We review the current state‐of‐the‐art of AI/ML‐guided CNS drug discovery, focusing on blood–brain barrier permeability prediction and implementation into therapeutic discovery for neurological diseases. Finally, we discuss the major challenges and limitations of current approaches and possible future directions that may provide resolutions to these difficulties.

INTRODUCTION

Disorders of the central nervous system (CNS) are responsible for multiple disease states of significant economic and social impact. Despite huge progress in our understanding of the structure and functions of the CNS, the development of new drugs for CNS disorders poses unique challenges. CNS drugs have lower success rates than other drug classes due to multiple factors, including an insufficient understanding of the pathophysiology of complex CNS conditions, poor target selection/engagement, lack of efficacy in early stages of development, and the presence of a blood–brain barrier (BBB). Such challenges have led to significantly longer development time for CNS drugs, which is, on average, 15–19 years to advance from discovery to regulatory approval. ¹ The whole process of developing a new drug generates a lot of data. Over the past decades, the advances in “omics” technologies, high‐throughput screening (HTS), and chemical synthesis have led to a dramatic increase in the amount of available data on chemical activity ² and functional genomics. ³ , ⁴ As a result, how to efficiently combine, correlate, and analyze existing large‐scale data has become a crucial problem for CNS drug discovery.

Artificial intelligence (AI) concepts such as machine learning (ML) have the potential to accelerate pharmaceutical research by extracting novel and important information from the vast amount of complex data generated from the drug discovery process. In recent years, AI/ML‐based methods have been widely applied to many therapeutic areas and achieved state‐of‐the‐art performance in addressing diverse problems in drug discovery. Such applications of AI/ML algorithms also have shown promise in the development of CNS therapeutics—the most challenging area in drug discovery. However, we have only just begun to explore the potential of these technologies for discovering novel therapeutics and repurposing old ones for CNS diseases. Therefore, this review will focus on AI/ML‐assisted drug discovery applications in this promising direction.

Here, we provide an overview of recent developments and applications in AI/ML‐assisted drug discovery, particularly for CNS diseases. This review is intended for biomedical researchers who are curious about the potential of AI/ML for advancing CNS drug discovery and consider AI‐based tools in their research. We first provide a broad overview of AI/ML approaches in drug discovery and then review AI/ML solutions to the issues in drug discovery specific for CNS diseases. We start with a brief introduction to AI algorithms and their input molecular descriptors and then summarize AI/ML‐based methods in various stages of drug discovery, including target identification and characterization, virtual screening, lead discovery, and physicochemical pharmacokinetic property prediction. We further review recent AI/ML applications in de novo design, predicting drug sensitivity and response, drug synergy prediction, and drug repurposing. For CNS diseases specific drug discovery, we focus on AI/ML solutions to key challenges such as BBB permeability and introduce AI/ML‐assisted applications to neurological diseases, including neurodevelopmental disorders, depression, Parkinson's disease, Alzheimer's disease, anesthesia and pain treatment. We conclude the review by highlighting challenges, limitations, and future directions of AI/ML‐aided drug discovery, especially for CNS diseases.

AI/ML APPLICATIONS IN DRUG DISCOVERY

AI/ML has been utilized at three different stages of early drug discovery process, including target identification, lead generation and optimization, and preclinical development (Figure 1). In target discovery, AI‐based approaches have been used to integrate heterogeneous data sets to identify patterns so as to understand molecular mechanisms underlying diseases and drug activities. For lead generation and optimization, AI/ML algorithms improve the scoring functions and quantitative structure–activity relationship (QSAR) models in virtual screening pipelines and support the automation and optimization of the de novo drug design processes. In preclinical development, AI/ML approaches are employed to generate predictive models of physicochemical properties by efficiently processing large amount of chemical data and further optimize absorption, distribution, metabolism, and excretion—toxicity (ADME‐T) profiles.

Figure 1 AI/ML applications in the drug discovery pipeline. AI/ML approaches provide a range of tools that can be applied in all the three stages of early drug discovery to improve decision making and speed up the process. ADME, absorption, distribution, metabolism, and excretion; AI, artificial intelligence; ML, machine learning; QSAR, quantitative structure–activity relationship (Figure not reproduced here; see the original article.)

Overview of the AI/ML algorithms

To help the reader better understand AI/ML applications in CNS drug discovery, we provide a summary of AI‐based algorithms that are widely used in drug discovery. AI uses a large variety of models to build up intelligent systems, which can be classified by learning procedures. AI is frequently used to denote ML algorithms—yet they are not the same. So, it would be worth clarifying both terms at first. In this review, we follow the US Food and Drug Administration's (FDA) definition of AI. They describe AI as “the science and engineering of making intelligent machines”, while ML is “an artificial intelligence technique that can be used to design and train software algorithms to learn from and act on data”, ⁵ adding that all ML techniques are AI techniques, but not all AI techniques are ML techniques. Here, we provide brief definitions of the basic learning algorithms in Table 1, as these are most relevant in the context of drug discovery. AI‐related learning techniques are broadly categorized as supervised, unsupervised, semisupervised, active, reinforcement, transfer, and multitask learning. Different algorithms are used in those learning architectures to perform specific tasks such as classification or clustering. However, success with AI requires more than training an AI model. A robust AI workflow involves (i) formulating a problem, (ii) preparing data, (iii) extracting features, (iv) selecting training and testing data sets, (v) developing a model, (vi) training the model and testing its performance (cross‐validation), and (vii) applying the model to testing data sets and refining the model. Figure 2 displays the basics steps of building an AI architecture.

Figure 2 The basic steps of building an artificial intelligence (AI) platform for drug discovery. The process for developing an AI model as follows: (1) Define the problem appropriately (objective, desired outputs, etc.), (2) prepare the data (collection, exploration and profiling, formatting, and improving the quality), (3) transform raw data into features and select meaningful features (a.k.a. feature engineering), (4) split data into training and validation sets, (5) develop a model, (6) train the model with a fraction of the data, test its performance (cross‐validation) and tune its parameters with the validation set (7) evaluate model performance on the validation set and refine the model, and (8) evaluate the model on independent data not used for method development (Figure not reproduced here; see the original article.)

Molecular descriptors and fingerprints for input data preparation

A key consideration in early drug discovery is to identify drug candidates with the desirable initial characteristics, which are then further developed into chemical structures with the desirable potency against the target molecule. Molecular descriptors and fingerprints are used for quantifying such physicochemical characteristics of both chemical entities and their biological target molecules. Molecular descriptors are experimentally quantified or theoretically characterized properties of a corresponding molecule that represent the physical, chemical, or topological characteristics, while molecular fingerprints are more complex descriptors that are encoded as binary bit strings. ⁶ , ⁷ Both molecular descriptors and fingerprints have crucial functions in ML‐based applications in drug discovery processes such as target molecule ranking, ⁸ , ⁹ similarity‐based compound search, ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ virtual screening, ¹⁶ , ¹⁷ QSAR analysis, ¹⁸ , ¹⁹ ADME‐T prediction of lead molecules. ²⁰ , ²¹ , ²² , ²³

There are various tools for molecular descriptor and fingerprint calculation, and each has a different set of features. Here, we explain the molecular descriptors (i.e., target protein descriptors and compound descriptors) and compound fingerprints, and provide the highly used programs for generating them (i.e., sequence‐based tools and structure‐based tools) in the Supporting Information. Additionally, Chuang et al. ²⁴ comprehensively discussed how AI‐based methods (i.e., deep learning [DL]) could address limitations of molecular descriptors and fingerprints and thereby improve the predictive modeling of compound bioactivities.

AI/ML applications in target identification

A dominant approach to drug discovery is to design drug molecules that will reverse a disease course by modulating the activity of a target. ²⁵ Drug development often begins with identification of a novel target whose modulation can lead to a therapeutic benefit with an acceptable safety margin. This is followed by validating the role of the selected target in disease in in vivo models and, ultimately, in clinical trials. Therefore, the ultimate success of a drug development project depends on early identification of promising drug targets.

A good drug target need be relevant to the disease phenotype as well as be suitable for therapeutic modulation (“druggable”). Biological and technological advances have continuously driven the generation of high‐throughput biomedical data, which present new opportunities for early identification of potential drug targets. However, the analysis of such large‐scale multidimensional biological data requires effective techniques that can produce accurate predictions for target identification. AI/ML has emerged as a powerful technology for analyzing the rapidly increasing multiomics data in the identification of potential therapeutic targets.

In literature, the “target identification” term is often used in two different contexts: Target discovery and target deconvolution. ²⁶ The first is the discovery of a new disease target whose modulation would have therapeutic effects. The second is the identification of a target with a known active compound, which is also called “target fishing.” To avoid confusion, we will use context‐specific terms of target discovery and deconvolution rather than generic target identification.

Target discovery

Drug discovery begins with the identification of a novel target candidate that is followed by a target evaluation consisting of experimental target validation and theoretical assessment of its ability to bind small molecule drugs (druggability). ²⁷ The target discovery process includes identification of targets that play a role in the disease pathophysiology, ²⁸ assessment of druggability, and prioritization of candidate targets. However, because of the complex nature of human diseases, this process often requires more comprehensive approaches that integrate available heterogeneous data and information to understand the molecular mechanisms underlying disease phenotypes and identifying the patient‐specific changes. ²⁹ To overcome such difficulties, researchers have applied AI/ML methods to predict “reliable” drug targets. The following sections demonstrate the AI/ML applications in different stages of the target discovery process (Figure 3).

Figure 3 AI‐guided target discovery. AI/ML methods can efficiently analyze all available information to speed up the discovery of disease‐related drug targets. Specifically, AI/ML methods are utilized for disease subtyping, identification of disease driver genes and microRNAs, alternative splicing prediction, triaging of novel drug targets, modeling of three‐dimensional target structures, and druggability assessment. AI, artificial intelligence; ML, machine learning (Figure not reproduced here; see the original article.)

Disease subtype prediction

In complex heterogeneous diseases, classifying patients into clinically and biologically homogenous subtypes is critical for understanding disease pathophysiology and developing appropriate subtype specific therapies. ³⁰ Researchers have developed AI/ML algorithms that can integrate multiscale data to identify different etiological subtypes of complex diseases. For example, Shen et al. ³¹ developed iCluster, a joint latent variable model for integrative clustering analysis, which was applied to breast cancer and lung cancer and identified subtypes characterized by concordant DNA copy number changes and gene expression. ³¹ Yuan et al. ³² also integrated copy number variation and gene expression data by using a nonparametric Bayesian model and discovered prognostic subtypes in prostate cancer and breast cancer. ³² Zhang et al. ³³ revealed the prognostic subtypes in neuroblastoma using DL‐based integration of multi‐Omics data and K‐means clustering analysis. Recently, Gao et al. ³⁴ described a cancer classification method, deep cancer subtype classification (DeepCC), based on DL of functional spectra, which is a vector of gene set enrichment scores associating with biological functions for each patient sample. Overall, in recent years, AI/ML methods have been employed to analyze large‐scale genomic and other molecular profiling data in cancer for the identification of distinct, molecular disease subtypes. However, such AI‐based subtyping analysis have not been widely applied to other complex diseases. Implementation of robust and scalable AI/ML techniques for discovery of disease subtypes paves the way for developing more efficacious therapeutic strategies.

Prediction of disease driver genes

One of the most challenging tasks in target discovery is the prediction of disease‐causing genes from huge amount of genetic and functional genomic data. To predict these disease‐associated genes from multiomics data, researchers have employed various ML classifiers, ³⁵ , ³⁶ , ³⁷ , ³⁸ including Random Forest (RF)‐, ³⁹ , ⁴⁰ support vector machines (SVM)‐, ⁴¹ , ⁴² and decision tree (DT)‐based classifiers. ⁴³ More detailed information about those applications can be found in the Supporting Information. Besides the ML‐methods using multiomics data, DriverML, ⁴⁴ a supervised learning tool, identified cancer driver genes based on DNA sequence alterations from The cancer Genome Atlas (TCGA) data with superior performance over the other tools such as DriverDBv2 database. ⁴⁵

In addition to ML classifiers, DL‐based methods have been implemented in more recently developed tools. For example, deepDriver ⁴⁶ trained similarity networks and a convolutional neural network (CNN) on mutation data simultaneously to predict driver genes with better performance than the competing approaches when applied in breast cancer and colorectal cancer. In another example, Peng et al. ⁴⁷ used deep neural network (DNN) to reduce the dimensionality of transcriptomics data to predict Parkinson's disease genes. This DNN‐based tool, namely, N2A‐SVM, consists of three steps, including extraction of vector representation of each gene in the protein–protein interaction (PPI) network, dimension reduction for the obtained vector with autoencoder, and prediction of the genes associated with Parkinson's disease using SVM.

Multitask learning has also been employed for the prediction of cancer driver genes. LOTUS, an ML‐based algorithm, predicts cancer driver genes in a pan‐cancer setting, as well as for specific cancer types, using a multitask learning strategy sharing information across cancer types. ⁴⁸ For the readers who want to learn more about opportunities and challenges in predictive modeling for multiomics data sets, we suggest the review paper of Kim and Tagkopoulos. ⁴⁹

Different from the tools using omics data sets, BeFree ⁵⁰ was developed to extract relations between genes and diseases from text mining. This supervised learning approach utilized natural language processing (NLP) Kernel methods to identify gene–disease associations from the abstracts collected by Medline.

Prediction of disease‐associated microRNAs

The challenges in targeting disease proteins have shifted the focus in target selection to disease microRNAs (miRNAs), which are small noncoding RNAs that regulate gene expression by targeting messenger RNAs. ⁵¹ miRNAs are regarded as high‐potential drug targets due to their involvement in various diseases. ⁵² Therefore, considerable effort has been devoted in identifying relationships between miRNAs and diseases using ML‐based methods, such as the network based approach by Xu et al. ⁵³ , ⁵⁴ and RLSMDA. New strategies in miRNA target discovery have utilized neural networks (NN). Zeng et al. ⁵⁵ developed a NN method, NNMDA to predict miRNA‐disease associations with the best performance among the existing algorithms. Application of NNMDA to lung neoplasm and breast neoplasm predicted novel disease‐related miRNAs. Very soon after that, Zheng et al. ⁵⁶ published a new ML‐based method, MLMDA, which predicts miRNA–disease associations by integrating miRNA sequence, disease semantics, miRNA–disease association, and miRNA function but with slightly worse performance than NNMDA.

Prediction of alternative splicing

Alternative splicing (AS) plays a fundamental role in gene expression regulation and protein diversity by causing the generation of different transcripts from single genes. ⁵⁷ Understanding the genetic variation in splicing signals is within the scope for AI/ML‐based models to discover therapeutic opportunities through novel targets. For splicing prediction and analysis, a web tool, AVISPA, ⁵⁸ has been developed. For a given exon and its proximal sequence, AVISPA predicts if the exon is alternatively spliced and if it has associated regulatory elements by using a Bayesian NN classifier. However, the method by Leung et al. ⁵⁹ outperformed the Bayesian NN approach for predicting AS by developing a DNN model inferred from mouse RNA‐Seq data that can predict splicing patterns in individual tissues and differences in splicing patterns across tissues. Later, Jha et al. ⁶⁰ compared those two previous modeling approaches, Bayesian and Deep NN, and determined the confounding effects of data sets and target functions. On the basis of this knowledge, they developed a new target function for AS prediction with higher accuracy. For further improvement of the prediction, they developed a modeling framework that uses transfer learning to combine CLIP‐Seq, knockdown, and overexpression experiments. For enabling the usage of unlabeled data and the latent information, Stanescu et al. ⁶¹ applied semisupervised learning algorithms to AS prediction. Xiong et al. ⁶² built up a DL model trained to predict splicing from DNA sequence alone and successfully identified new autism‐linked genes.

Target prioritization

While increasing effort has been devoted to nominating novel drug targets involved in diseases, experimental validation of identified target candidates is an expensive and time‐consuming task. ⁶³ Therefore, researchers have utilized AI/ML approaches to support the prioritization of the most promising target candidates for subsequent experiments. To identify and prioritize novel cancer drug targets, Jeon et al. ⁶⁴ built an SVM classifier that uses features from various data types (DNA copy number, messenger RNA expression, mutation occurrence, and PPI) to prioritize drug targets specific for breast, pancreatic and ovarian cancers. To improve the disease gene prioritization process, Valentini et al. ⁶⁵ combined different functional gene networks and applied a kernel‐based method to prioritize genes according to the disease MeSH terms. Then, Ferrero et al. ⁶⁶ took advantage of the publicly available target–disease association data from the open targets platform training an NN classifier with semisupervised learning and predicted novel therapeutic targets. As another publicly available data source, Medline abstracts also have been benefited for developing prediction tools (i.e., DigSee ⁶⁷ ) that identify disease–gene relationships and prioritize the genes based on evidence. Specifically, DigSee uses NLP to extract the relationship between diseases and genes and ranks the evidence sentences with a Bayesian classifier. Recently, Arabfard et al. ⁶⁸ predicted and prioritized over 3,000 candidate age‐related human genes using three positive unlabeled learning algorithms, Naïve Bayes, Spy, and Rocchio‐SVM. They ranked the human genes according to their implication in aging based on binary gene features from 11 human biology databases. ⁶⁸

Target protein structure prediction

AI/ML architectures have been applied in protein structure prediction over 30 years, and several groups have comprehensively reviewed those strategies. ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ Therefore, we will focus on recent applications in this field. Also, we provide a background of conventional protein structure prediction methods (i.e., template‐based and template‐free) for those who want to learn more about this field in the Supporting Information.

Since 1994, the Critical Assessment of protein Structure Prediction (CASP) competitions have been organized biannually for blind evaluation of the state‐of‐the‐art methods that predict three‐dimensional (3D) protein structures from protein sequences. There, each group submits structure predictions for each of the given protein sequences for which experimentally determined structures were sequestered. In December 2018, Google's AI firm DeepMind won the CASP13 competition with its latest AI system, AlphaFold. DeepMind's success generated significant interest in the protein folding community, where the researchers published several articles discussing the method. ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ AlphaFold determines the 3D shape of a protein from its amino acid sequence by merging two approaches: (i) Inferring physical contact in protein structure from residue covariation in protein sequence based on coevolution analysis of a multiple sequence alignment and (ii) identifying coevolutionary patterns in protein sequences as contact distributions by using DNNs and convert them into protein‐specific statistical energy potentials. AlphaFold system has achieved an unprecedented prediction accuracy among the ab initio methods. Although AlphaFold's performance represents a big leap in protein structure prediction, its accuracy still needs to be improved.

Inspired by AlphaFold as well as previous successful applications of DL to residue contact predictions, ⁷⁸ researchers have developed different strategies to improve the protein structure prediction, including a deep residual network model, ⁷⁹ a fragment library that is built using deep contextual learning techniques called DeepFragLib ⁸⁰ and a community‐built, open‐source implementation of Alphafold (i.e., ProSPr). ⁸¹ The emergence of DL has suggested the rethinking of how to address the problem of protein structure and thereby, encourages the new approaches. RGN (recurrent geometric network) is an end‐to‐end differentiable model that takes a sequence of amino acids and position‐specific scoring matrices (a summary of residue propensities for mutation) as inputs and outputs a 3D structure. In contrast to the complexity of conventional structure prediction models, a trained RGN model is a single mathematical function that is evaluated once per prediction. Hence, a trained RGN makes predictions six to seven orders of magnitude faster than other methods. The same lab developed the RGN also published a data set to provide a standardized resource for training and assessing ML frameworks for predicting protein structures. The data set called ProteinNet integrates sequence, structure, and evolutionary information into preformatted input/output records. ProteinNet is available in a public repository, https://github.com/aqlaboratory/proteinnet.

Going beyond the structure prediction, researchers have employed the ML for the prediction of protein dynamics since target proteins are dynamic and sample multiple states. Ung et al. ⁸² used RF to classify pharmacologically relevant conformations of protein kinases. Using a 3D‐CNN, Okuno et al. ⁸³ developed DEFMap, which extracts the dynamics information hidden in a given cryo‐EM density map. This approach allows us to grasp the dynamic changes associated with molecular recognition and the accompanying conformational selections from the cryo‐EM structure, which derive insights into the protein function as well.

The studies discussed above clearly demonstrate the utility of the AI/ML frameworks to make predictions of protein structural features from sequence alone. Rost et al. ⁸⁴ comprehensively discussed how ML algorithms help to understand the effects of protein sequence variants on protein function and pathways. AI/ML algorithms are readily available for structural biologists to quickly estimate protein structures. Of course, the accuracy and speed of a framework will depend on the creativity in problem formulation, network design, and data storage. We can look forward to a rapid growth in the number of AI/ML applications in the prediction of protein structures.

Druggability

In target discovery, another crucial step is the evaluation of the target's druggability, “the likelihood of being able to modulate a target with a small‐molecule drug”. ⁸⁵ In drug design, a selected target must have the biophysical properties that allow it to bind small molecules with drug‐like properties. ML‐based models usually estimate a target's druggability by using different features of it. As one of the earliest applications, SCREEN (Surface Cavity REcognition and EvaluatioN) webserver ⁸⁶ was built based on an RF classifier trained on geometric, structural, and physicochemical features of drug‐binding and nondrug‐binding cavities on proteins. The classification process reveals that the most critical attributes to estimate druggability are the size and shape of the surface cavities of the protein. In the following studies, SVMs were applied to predict druggable targets based on various physicochemical properties from protein sequences. ⁸⁷ , ⁸⁸ Then, Costa et al. ⁸⁹ constructed a DT‐based meta‐classifier by training on attributes including network topological features, tissue expression profile, and subcellular localization for each druggable and nondruggable gene. Later, Wang et al. ⁹⁰ combined a biased SVM with a DL model, stacked autoencoders, to identify drug target proteins based on the sequence information of proteins. Recently, Kokh et al. ⁹¹ developed an ML tool for the druggability analysis of binding pocket variations during the protein movement. They used a logistic regression model and a CNN to identify potentially druggable protein conformations in trajectories from molecular dynamics simulations. On the contrary, Dezső and Ceccarelli ⁹² built up RF models for the druggability prediction of oncology drug targets to prioritize proteins according to their similarity to approved drug targets. More details on ML‐based tools designed to predict the druggability of targets can be found in the review from Kandoi et al. ⁹³

Target deconvolution

Target deconvolution (a.k.a. target fishing) is an important step following the discovery of compounds that cause a desirable change in phenotype. Understanding the binding targets of phenotypic screen‐derived compounds can help design better analogs, find potential off‐targets, and thereby explain observed adverse events. However, existing experimental approaches for target deconvolution are labor, resource, and time‐intensive. Researchers have adapted computational approaches to target deconvolution problems to reduce the required sources for the experiments. Several studies implemented AI/ML algorithms into computational target deconvolution tools for higher predictive power. For example, Schneider and colleagues have widely applied self‐organizing maps (SOMs) to predict the macromolecular targets of compounds. ⁹⁴ , ⁹⁵ , ⁹⁶ , ⁹⁷ They preferred to use “fuzzy” molecular representations, such as pharmacophoric feature descriptors, since such fuzzy molecular representations demonstrated greater scaffold‐hopping potential than atomistic approaches in similarity searches. On the basis of the similarity of pharmacophoric features, their unsupervised SOM algorithm clustered the query molecules with unknown targets as well as drug‐like molecules with known targets. Hence, the trained SOM was able to transfer the knowledge of annotated drug targets to query molecules that are the nearest neighbors to known drugs. ⁹⁴ They have applied this SOM approach to identify the macromolecular targets of de novo‐designed molecules, ⁹⁵ complex natural products, ⁹⁴ fragment‐like natural products, ⁹⁶ and a natural anticancer compound. ⁹⁷ Besides the SOM models, a multiple‐category Naïve Bayesian model was developed for the rapid identification of potential targets for compounds based on only chemical structure information, which is the connectivity fingerprints of compounds from 964 target classes in the WOMBAT (World Of Molecular BioAcTivity) chemogenomics database. ⁹⁸ Moreover, a target‐fishing server named RF‐QSAR was built based on target SAR models that were created using an RF algorithm to rank candidate targets for a query compound. ⁹⁹ A recent target identification tool, BANDIT, ¹⁰⁰ uses a Bayesian approach to integrates six distinct data types—drug efficacies, posttreatment transcriptional responses, chemical structures, reported side effects, bioassay results, and known targets.

In the identification of the novel targets of drugs, there has been increasing interest in predicting drug–target interaction (DTI), given its relevance for side effect prediction and drug‐repositioning attempts. ¹⁰¹ The availability of heterogeneous biological data on known DTI has enabled the development of various AI/ML‐based strategies to exploit unknown DTI, ¹⁰² including ensemble learning, ¹⁰³ , ¹⁰⁴ , ¹⁰⁵ , ¹⁰⁶ tree‐ensemble learning, ¹⁰⁷ active learning, ¹⁰⁸ DL, ¹⁰⁹ end‐to‐end DL, ¹¹⁰ and kernel‐based learning. ¹¹¹ , ¹¹² , ¹¹³ , ¹¹⁴ , ¹¹⁵ Such AI/ML‐enabled data integration strategies outperform the traditional methods in classifying both positive and negative interactions, ¹¹⁰ improved the quality of the predicted interactions, and expedited the identification of new DTI. ¹¹⁵

AI/ML applications in compound screening and lead discovery

To identify new compounds with potential interactions to target proteins, researchers commonly use HTS, an in vitro method that automatically tests large compound libraries towards a specific target. However, high cost and low hit rate of HTS have expedited the development of virtual screening (VS) alternatives, which enable cheaper and faster screening of larger compound libraries. ¹¹⁶ , ¹¹⁷ VS predicts the compounds that most likely to bind to a protein of interest using various approaches. Two broad categories of VS are structure‐based VS (SBVS) and ligand‐based VS (LBVS)—the former takes the structures of target proteins as input, ¹¹⁸ , ¹¹⁹ and the latter uses information on known inhibitors. ¹²⁰ LBVS is basically “analoging” to some extent based on that similar molecules tend to exhibit similar properties, ¹²¹ and it also helps to build better pharmacophore models. SBVS and LBVS are often used synergistically: Leads from SBVS can be improved with LBVS, and data from improved yields can be used to refine models for SBVS. ¹²² For achieving better performance in VS workflows, AI/ML‐based methods have been utilized for both SBVS and LBVS. We will begin with the application of AI/ML methods in SBVS and continue with their applications in LBVS in the next section.

Structure‐based virtual screening

SBVS requires the 3D structure of a target protein to predict whether a compound is likely to bind the target. One widely used method to do this is molecular docking, which models the protein–ligand complex based on the estimated interaction energy. In recent years, ML methods have been employed in SBVS workflow to increase the robustness and accuracy of scoring functions (SFs), conformational sampling and ranking. Researchers have developed SFs using RF‐, ¹²³ , ¹²⁴ , ¹²⁵ , ¹²⁶ SVM‐, ¹²⁷ , ¹²⁸ and NN‐ ¹²⁹ , ¹³⁰ , ¹³¹ , ¹³² , ¹³³ , ¹³⁴ based learning algorithms and they outperformed the conventional SF predictions. ¹³⁵ However, no ML‐based SF is superior to all the other approaches in all respects. ¹³⁶ Indeed, the performance of an SF differs from target to target. ¹³⁷ Therefore, researchers have developed ML‐based, target‐specific SFs to improve the efficiency of existing SFs for kinases, ¹³⁸ , ¹³⁹ , ¹⁴⁰ , ¹⁴¹ histone methyltransferases, ¹⁴² cyclin‐dependent kinases and G protein‐coupled receptors (GPCRs), ¹³⁷ and cytochrome P450 aromatase. ¹⁴³

Moreover, such ML‐based models have been applied to post‐docking processes to improve the accuracy of molecular docking. For example, ML algorithms ¹⁴² , ¹⁴⁴ , ¹⁴⁵ , ¹⁴⁶ , ¹⁴⁷ , ¹⁴⁸ improve pose/compound selection by automating the evaluation of docked ligands, which was done manually before. ¹⁴⁹ Details about ML‐based scoring functions and AI/ML applications in the post‐docking stage can be found in the Supporting Information.

Ligand‐based virtual screening

When the 3D structure of a given target is available, SBVS approaches (i.e., molecular docking) can be employed. However, LBVS methods are the only option if the 3D structure of the target protein is not known. In contrast to the molecular docking that predicts the binding pose of ligands to the target protein using the protein structure, LBVS is based on the principle that ligands structurally similar to an active compound tend to have similar activity. ¹⁵⁰ Hence, LBVS requires the information of known active compounds rather than the target protein structure. In drug discovery efforts, researchers often have a set of active compounds generated from testing molecules in biochemical or functional assays without knowing the target protein structure. In such cases, the LBVS approach can be utilized to find new ligands by assessing the structural similarity of candidate ligands to the known active compounds. The challenge is thereby to find an appropriate model for similarity that relates compound features to assay outcomes. In recent years, ML has emerged as an attractive approach to boost the predictive power of LBVS models. The specific aims of ML approaches include prediction of the active compounds against a particular target using models trained on input data sets, discrimination of drug modules from nondrug ones, and prioritization of compounds based on the probability of activity. For these purposes, researchers have used SVMs, Bayesian architectures, and artificial neural networks (ANNs) (Table S2). Further information regarding AI/ML applications in LBVS is available in some comprehensive review papers. ¹³⁶ , ¹⁵¹ , ¹⁵²

On the contrary, one of the most recent advances in AI/ML‐based LBSV was made by Stokes et al. ¹⁵³ They successfully discovered new antibiotics by employing graph convolutional networks (GCN), whose outstanding performance over conventional ML models in predicting molecular properties was confirmed by two studies. ¹⁵⁴ , ¹⁵⁵ Using their GCN model, the authors performed a large‐scale screening and identified a promising new antibiotic, halicin. ¹⁵³

In conclusion, the advances in selection and design of AI/ML algorithms for LBVS and the availability of large bioactivity data sets have enabled more accurate and faster selection of compounds that are predicted to be active against a particular target and will undergo further experimental assays eventually. Although traditional ML classifiers had been widely used in LBVS, recent successful applications have shown GCN's potential to become a popular approach for LBVS. ¹⁵¹

QSAR prediction

QSAR models are developed to identify a mathematical relationship between the physicochemical properties, which are represented by molecular descriptors, and biological activity of chemicals. These models play a prominent role in drug optimization, providing a preliminary in silico evaluation of essential attributes related to the activity, selectivity, and toxicity of candidate compounds. ¹⁵⁶ , ¹⁵⁷ , ¹⁵⁸ By doing that, they significantly reduce the number of candidate compounds to be tested by in vivo experiments. QSAR models can be based on regression or classification models that depend on the underlying computational strategy. AI/ML approaches (i.e., RF, ¹⁵⁹ , ¹⁶⁰ SVM, ¹⁶¹ , ¹⁶² , ¹⁶³ Naïve Bayesian, ¹⁶⁴ , ¹⁶⁵ , ¹⁶⁶ , ¹⁶⁷ , ¹⁶⁸ , ¹⁶⁹ , ¹⁷⁰ , ¹⁷¹ , ¹⁷² , ¹⁷³ and ANN ¹⁴³ , ¹⁷⁴ , ¹⁷⁵ , ¹⁷⁶ , ¹⁷⁷ , ¹⁷⁸ , ¹⁷⁹ , ¹⁸⁰ , ¹⁸¹ , ¹⁸² , ¹⁸³ , ¹⁸⁴ ) have been extensively employed in QSAR modeling (For the detailed discussion of the applications, see the Supporting Information). Notably, the RF algorithm is commonly used as a classification and regression tool ¹⁵⁹ and considered to be the golden standard in QSAR studies. ¹⁸⁵ Hence, the performance of new QSAR prediction tools often is compared with that of RF. Many RF‐based QSAR models have been developed, such as pQSAR, ¹⁸⁶ a method for the soluble epoxide hydrolase, ¹⁸⁷ and a model for Janus kinase 2. ¹⁸⁸ When the predictive performance and interpretability of RF‐based QSAR models are compared to those of two widely used linear modeling approaches—SVMs and partial least‐squares, RF not only yields better predictive performance but also enables an amenable chemical and biological interpretation. ¹⁸⁹

In the applications of NN to QSAR prediction, researchers use the data from a single assay using molecular descriptors as input to train an NN and record activities as training labels. However, the efficiency of those simple single‐task NN models depends on having sufficient training data in a single assay. To benefit from the data obtained from multiple assays, researchers aim to develop multitask QSAR models. Several groups constructed the multitask learning structures based on plain feed‐forward NN to avoid overfitting by learning multiple bioassays simultaneously. ¹⁹⁰ , ¹⁹¹ , ¹⁹² , ¹⁹³ , ¹⁹⁴ , ¹⁹⁵ , ¹⁹⁶ Moreover, multitask QSAR models were also utilized for predicting the activity against multiple targets. ¹⁹⁷ , ¹⁹⁸ , ¹⁹⁹

In 2012, a data science competition (www.kaggle.com/c/MerckActivity) was organized to find state‐of‐the‐art methods for QSAR. Using multitask DNNs, the winning team improved the prediction accuracy by 15% over the baseline RF method. ²⁰⁰ Since its introduction into the QSAR modeling, ¹⁵⁹ RF has served as a “golden standard” and no QSAR methods other than DNNs outperform it. On the contrary, in the following DREAM challenges on predicting kinase‐drug‐binding, ²⁰¹ the models based on DL algorithms did not perform better than the other learning algorithms. ²⁰² In the next study, using the DNNs, Ma et al. ¹⁸⁵ showed that DNNs could make better prospective predictions than RF, on large and diverse QSAR data sets. However, they could not propose a clear strategy for choosing between multitask and single‐task DNNs. Xu et al. ²⁰³ focused on demystifying multitask DNNs and explored why multitask DNNs perform significantly better or worse for some QSAR tasks. They found that multitask DNNs can boost the predictive performance if the assistant tasks have molecules in a training set with structures similar to those in the test set of the primary task and the activities between these similar molecules are correlated. Contrarily, if the assistant tasks do not include compounds structurally similar to those in the primary task test set, multitask DNNs show no improvement in prediction, regardless of correlated or uncorrelated activities. Recently, Zakharov et al. ²⁰⁴ combined multitask DNNs with consensus modeling to generate large‐scale QSAR models with improved prediction accuracy over the state‐of‐the‐art QSAR models.

Ensemble‐based ML approaches combining several basic models have also been used to overcome the weaknesses of individual learning models and thereby improve the overall performance of the QSAR predictors. There are various ensemble learning applications in QSAR predictions, including data sampling ensembles, method ensembles, and representation ensembles. Recently, Kwon et al. ²⁰⁵ proposed a model that is a combined ensemble of sampling, method, and representation with an end‐to‐end NN‐based individual classifier. Their ensemble model achieved better performance than the individual models in QSAR prediction.

AI/ML applications in prediction of physicochemical properties and ADME‐T

Prediction of physicochemical properties

Physicochemical properties indicate all aspects of drug action and profoundly affect the clinical success rates of drug candidates. A small molecule drug candidate must be sufficiently soluble and permeable to access its site of action and thereby engage its targets, with optimal safety profiles. Therefore, accurate prediction of the physicochemical characteristics can be beneficial for designing a new chemical entity with suitable pharmacokinetic and pharmacodynamic profiles. Researchers have adopted ML‐driven approaches to predict some key physicochemical properties, such as water solubility, membrane permeability, and lipophilicity. We provide a detailed description of each property and discuss the ML‐based techniques that specifically predict the water solubility, ²⁰⁶ , ²⁰⁷ , ²⁰⁸ , ²⁰⁹ , ²¹⁰ membrane permeability, ²¹¹ , ²¹² , ²¹³ and lipophilicity ²¹⁴ , ²¹⁵ , ²¹⁶ , ²¹⁷ , ²¹⁸ , ²¹⁹ in the Supporting Information. Although improved ML models have led to better prediction of molecular properties, the lack of standard criteria for performance evaluation has limited the progress. To address this, MoleculeNet, a benchmark collection for molecular ML was developed to serve as a unique resource for the scientific community to create advanced models for learning molecular properties. ¹⁵⁴ To further support the comparison and development of novel models, MoleculeNet has implemented various ML algorithms. Benchmark results have shown that graph convolutional network (GCN) outperforms other traditional ML methods based on molecular fingerprints and descriptors to predict molecular properties. Recent studies have supported the superior performance of GCN. Applying GCN, Feinberg et al. ¹⁵⁵ achieved an unprecedentedly high accuracy in predicting molecular physicochemical properties.

ADME‐T predictions

A successful drug development pathway must include the evaluation and optimization of pharmacokinetics, pharmacodynamics, and safety profiles of a candidate molecule. In early drug discovery, evaluation of the ADME‐T properties help researchers select good drug candidates for further development. ADME‐T properties are estimated to be responsible for half of all clinical failures. ²²⁰ In this context, in silico ADME‐T prediction models have received considerable progress over the past 40 years due to the availability of many compounds with known pharmacokinetic properties. ²³ , ²²¹ Prediction models usually try to build a direct relationship between a set of molecular descriptors and a given ADME‐T property. ²²² These methods represent a compound by chemical descriptors as input features such as atom counts, surface areas, weight, van der Waals volume, partial charge information, and the presence or absence of a predefined substructure. The key substructures responsible for certain toxicity are structural alerts, of which detection in given small molecules could be used for toxicity prediction. ²²³ On the contrary, in these models, the toxicity properties of input compounds are HTS assay measurements of toxic effects that are highly relevant to human health, including nuclear receptor pathway assays (i.e., aryl hydrocarbon receptor, aromatase, androgen and estrogen receptor, PPAR‐gamma) and stress response pathway assays (i.e., ATAD5, antioxidant responsive element, heat shock factor response element, mitochondrial membrane potential, p53). ²²⁴ While the conventional approaches have yielded physiologically based pharmacokinetic and pharmacokinetic‐pharmacodynamic/quantitative systems pharmacology models, researchers have applied AI/ML algorithms to produce high‐quality models with improved accuracy and thus provide meaningful predictions of ADME‐T responses using chemical structure information. For predicting regulators of drug ADME‐T properties, the classification models—DT, K‐nearest‐neighbor (KNN), SVM, RF, and NN have been extensively used. Even beyond that, the introduction of DL models has led to further developments in this area. As a good example of recent advancements in AI. ML‐aided ADME‐T prediction, Alchemite ²²⁵ —a DL model—predicts ADME‐T properties by imputing heterogeneous drug discovery data, including multitarget biochemical activities, phenotypic activities in cell‐based assays, and ADME‐T endpoints.

Moreover, the introduction of capsule networks, a new class of DNN architectures, has remarkably improved the ADME‐T prediction. To predict the cardiotoxicity of drugs, Wang et al. ²²⁶ developed two capsule network architectures, including a convolution‐capsule network (Conv‐CapsNet) and a restricted Boltzmann machine‐capsule network (RBM‐CapsNet). Both models showed excellent performance with an accuracy of 91.8% for Conv‐CapsNet and 92.2% for RBM‐CapsNet. As the volume and chemotype coverage of the available ADME‐T databases are continually growing, we have witnessed a great progress in AI/ML‐guided ADME‐T prediction in recent years. Such advances in the field have been extensively reviewed. ¹³⁶ , ²²⁷ , ²²⁸ , ²²⁹ , ²³⁰ , ²³¹ , ²³²

AI/ML applications in de novo drug design

In de novo drug design, scientists generate novel chemical entities with desired chemical and biological characteristics from scratch, aiming to achieve particular efficacy and safety profiles in a cost‐ and time‐efficient manner. Advanced AI/ML‐based tools have enabled the automated generation of new chemical entities with suitable properties. As a result of such achievements, application of AI/ML to de novo discovery has become a popular topic over the last few years. Particularly, generative molecular design based on AI/ML has aroused considerable attention. In this section, we summarize the AI/ML algorithms utilized for de novo drug design with a focus on generative models. Those who want to learn more about this subject can check other comprehensive sources in the literature. ¹³⁶ , ²³³ , ²³⁴

Traditional methods for generating novel chemical structures depend on the previously defined reaction or transformation rules, which bias the chemical space towards prior chemical knowledge. AI/ML‐based generative models are entirely data‐driven without relying on any explicit rules and can generate new molecules that are not present in a training set. Briefly, these generative models first learn from data, then create an abstract representation of the data, and finally use this representation to generate new data instances. ²³⁵ Thus, these generative models demonstrate all aspects of an artificially intelligent system (i.e., problem‐solving, learning from experience, and coping with new situations). ²³⁵

Recent de novo molecule‐generative models with an ML structure include adversarial autoencoders (AAE), ²³⁶ , ²³⁷ , ²³⁸ variational autoencoders (VAE), ²³⁹ , ²⁴⁰ and recurrent neural networks (RNN). ²⁴¹ , ²⁴² , ²⁴³ , ²⁴⁴ In generating novel molecules represented by simplified molecular input‐line entry system (SMILES) strings, RNN is a promising approach for learning from large sets of SMILES strings and generating ligands with similar activities to those of the training set templates, but with novel scaffolds. However, the percentage of valid SMILES, internal diversity, and the similarity of molecules to the training data set in the libraries generated by any given approach have been a matter of debate. To address these issues, Reinforcement learning (RL) has been embedded in ML architectures. ²³⁹ , ²⁴⁵ , ²⁴⁶ , ²⁴⁷ , ²⁴⁸ Introducing a task‐specific reward function, RL‐assisted models are able to produce chemically feasible and predominantly novel molecules with appropriate molecular properties. For the generation of novel small molecules with the desired characteristics, generative adversarial networks (GAN) also have been employed. For example, druGAN ²³⁷ (drug‐generative adversarial network) has been developed for producing new molecules with specific anticancer properties.

Another commonly used drug design approach is to generate new analogs/similar drugs of a given set of drugs. In such cases, the transfer learning models have been integrated into NN architectures to increase the prediction accuracy by taking knowledge acquired from training on a previous problem and applying them to a new but related problem. ²⁴⁹ , ²⁵⁰

In the generative drug design models above, many ML architectures use the SMILES as molecular representation. SMILES provides a linear representation, referred to as a SMILES string that can be translated into a graph and enables a straightforward application. However, it has one or more limitations: Generated SMILES may not represent a chemically feasible structure, and even a single character alteration in a SMILES representation can change the underlying molecular structure significantly. ²⁵¹ To overcome its limitations, researchers proposed several solutions like converting SMILES strings into a new SMILES‐like syntax ²⁵² or utilizing grammatical evaluation of the SMILES syntax. ²⁵³ Besides the SMILES string representation, molecular graphs have also been used to train ML‐based molecule generation algorithms. ²⁵⁴ In molecular graph generators, structures are directly represented as graphs in every step and substructures are inferred from the partially generated molecular graphs. ²⁵⁵ Examples of such ML models to design de novo molecules based on graph representation includes GANs ²⁵⁶ , ²⁵⁷ and VAEs. ²⁵⁸ , ²⁵⁹

In addition to the models mentioned above, some AI/ML‐driven de novo molecule design tools are distinguished by introducing novel approaches. An automated de novo molecular design tool, DINGOS, ²⁶⁰ has been developed to emulate the approach of a synthetic chemist. It assembles drug‐like new compounds through modular and synthetically feasible design schemes, considering the synthetic feasibility of each step. In brief, the DINGOS algorithm combines a rule‐based approach with an ML model trained on known successful synthetic routes, while the former ensures the synthesizability and the later provides a directed approach to limiting the output molecules to compounds with desirable similarity to the template. Another remarkable ML‐based generative approach is proposed by Méndez‐Lucio et al., ²⁶¹ which bridges systems biology and molecular design. To our knowledge, it is the first AI/ML‐based drug design tool that combines transcriptomic and structural data. Conditioning a GAN architecture with compound‐induced transcriptomic data (i.e., L1000 data set), they can automatically design molecules that potentially produce the desired transcriptomic outcome. Their model allows the design of active‐like molecules for a desired target using just gene expression signature of target perturbation. However, the current version is not capable of generating compounds that can reverse disease‐related gene expression signatures. Also, its performance has not been evaluated in a real drug‐discovery setting yet.

Among all the studies of AI/ML‐based generative molecular design, maybe the most‐mentioned ²⁶² one is published by Insilico Medicine, ²⁶³ showing how AI for generative chemistry can be used to drive rapid drug discovery. The goal of the study was to demonstrate that efficacious drugs can be developed in just 21 days for a new target. For this purpose, they have developed a generative tensorial reinforcement learning (GENTRL) model, which can be seen as an advanced version of their earlier algorithms on VAE ²³⁸ and GAN, ²⁴⁷ to design DDR1 kinase inhibitors. Notably, this study has two major limitations: First, DDR1 is considered to be the most promiscuous kinase ²⁶⁴ ; thus, developing compounds targeting this protein may be considered low hanging‐fruit. Second, the seemingly novel compound is highly similar to the widely used cancer drug ponatinib, indicating the limitation of the approach ²⁶⁵ in assessing truly novel scaffolds. Therefore, there is still room for improvement of AI/ML‐inferred small molecules to obtain a clinical candidate.

AI/ML applications in prediction of drug sensitivity and response

Personalized drug response prediction aims to improve the targeted therapy response in complex diseases like cancer. ²⁶⁶ However, the limited application of candidate drugs in clinical settings and the heterogeneity among cancer patients make it difficult to tailor therapy for each individual cancer patient. Personalized treatment design requires predictive methods that are capable of exploiting large, heterogeneous, and sparsely sampled data sets. Accurate AI/ML‐based models employing in vitro and in vivo data sets have the potential to improve the prediction of response of cancer cells to a given compound. There are various AI/ML models to predict drug sensitivity and anticancer drug response. In such efforts, elastic net regression, ²⁶⁷ , ²⁶⁸ , ²⁶⁹ ensemble‐based approaches, ²⁷⁰ , ²⁷¹ transfer learning, ²⁷² autoencoders, ²⁶⁶ , ²⁷³ , ²⁷⁴ , ²⁷⁵ and multitask learning approaches ²⁷⁶ , ²⁷⁷ , ²⁷⁸ have been widely used. The details about these AI/ML applications can be found in the Supporting Information.

AI/ML applications in prediction of drug–drug interactions

In the treatment of complex diseases such as neurological disorders, diabetes, cancer, or cardiovascular disease, drug combinations are highly utilized for medical intervention. Coadministration of drugs in the treatment aims to enhance efficacy, reduced toxicity, and prevent the emergence of resistance. Drug combinations are classified as synergistic, antagonistic, or additive. Drug synergy is the interaction of two or more drugs, causing the total effect of drugs to be greater than sum of individual effects of each drug. ²⁷⁹ If drugs act synergistically, lower doses of each drug could potentially be enough to provide the desired outcome allowing for less adverse effects. Opposite to synergism, the antagonistic combination means that the combined activity of the drugs is lower than the response of the individual agents. ²⁸⁰ Finally, a drug combination is considered to be additive when the response of each drug neither masks nor enhances the efficacy of others. ²⁸¹ Although combinatorial therapy has advantages over monotherapy, developing a new drug combination regimen that can be transferred to the clinic is still challenging. So far, the effective drug combinations have been suggested based on either clinical experience or HTS of drug pairs at different concentrations on cell lines. However, the former involves the risk of harm to patients, and the latter is unfeasible to test the complete combinatorial space. ²⁸² To accelerate conventional combinatorial therapy efforts, AI/ML algorithms have begun to be utilized for prioritizing the drug pairs and exploring the larger combinatorial space. Tonekaboni et al. ²⁸³ introduced some examples of various ML‐based prediction frameworks for drug–drug interactions. To avoid duplication, we overview the AI/ML applications in combinatorial therapy after that time, including the applications in cancer ²⁸⁴ , ²⁸⁵ , ²⁸⁶ , ²⁸⁷ , ²⁸⁸ and depression treatment, ²⁸⁹ antimalarial, ²⁹⁰ and antibiotic ²⁹¹ discovery, along with the available AI/ML‐based tools to predict the synergistic effects of drug combinations ²⁹² , ²⁹³ , ²⁹⁴ in the Supporting Information.

In addition to the synergistic effects, drug–drug interactions can induce unexpected adverse drug reactions. Such adverse reactions caused by drug–drug interactions could lead to death in some extreme cases. ²⁹⁵ Therefore, AI/ML‐based models have been developed to predict the risk of side effects due to drug–drug interactions. Applications of GCN, ²⁹⁶ DNN, ²⁹⁷ and ML architectures ²⁹⁸ showed promising results for predicting adverse drug reactions of drug combinations. Lee and Chen ²⁹⁹ extensively discussed the role of ML approaches in detection and classification of side effects caused by drug–drug interactions in their review of previous studies. In a recent study, Shankar et al. ³⁰⁰ predicted the adverse drug reactions of coadministered drug pairs using an ANN trained on transcriptomic data, compound chemical fingerprint, and Gene Ontologies. ³⁰⁰

AI/ML applications in drug repurposing

Drug development and trials in animals and humans is a time‐consuming and expensive process. In general, the whole process for developing a new FDA‐approved drug requires 10–17 years of period and the tremendous cost of $2.6 billion. ³⁰¹ However, high expenditures for drug development has not been able to increase the rate of approved drugs. ³⁰² Among the reasons for this limited approval rate, a key factor is the continued adherence to the classical “one gene, one drug, one disease” paradigm in the traditional drug development. ³⁰³ Since drug targets do not operate in isolation from the biochemical system, each DTI must be studied in a broader integrative context. ³⁰⁴ This approach provides new insights into “off‐target” effects (i.e., side effects), resistance to precision therapy, and drug mechanism of action that can inform drug‐repurposing efforts.

Drug repurposing, also known as drug repositioning, denotes the new indications of existing drugs and is an alternative over the de novo drug development. Although the unknown underlying complex biology and pharmacology has challenged the drug‐repurposing attempts, intelligent computer algorithms offer a strategy for detecting potential drug indications by integrating large‐scale heterogeneous data (i.e., genomic, transcriptomic, phenotypic, chemical, and bioactivity) from hundreds of approved drugs. Various specially designed AI/ML models have been proposed for detecting novel drug indications. Here, we classify the ML applications for drug repositioning into the following three categories: (i) Similarity‐based methods that employ different types of classifiers like logistic regression, ³⁰⁵ , ³⁰⁶ SVM, ³⁰⁷ , ³⁰⁸ , ³⁰⁹ RF, ³¹⁰ , ³¹¹ KNN, ³¹² and CNN, ³¹³ (ii) feature vector‐based methods that utilize supervised ³¹⁴ , ³¹⁵ , ³¹⁶ , ³¹⁷ , ³¹⁸ and semisupervised ³¹⁹ , ³²⁰ , ³²¹ learning algorithms, and (iii) network‐based methods that mainly use semisupervised learning algorithms (e.g., Laplacian regularized least square, ³²² , ³²³ , ³²⁴ label propagation, ³²⁵ random walk, ³²⁶ and RF ³¹⁰ ). We provide an in‐depth discussion of these three classes of AI‐based drug repositioning applications in the Supporting Information. Particularly, in early 2020, researchers at MIT published a milestone paper using a DL approach to antibiotic discovery. ¹⁵³ They trained the deep GCN model based on molecular features and predicted halicin as an antibacterial molecule from the Drug‐Repurposing Hub. Halicin showed a broad‐spectrum activity against drug‐resistant strains in mice. This is the first time an AI/ML‐assisted tool was used to identify thoroughly new types of antibiotic from scratch, without the need for any previous human assumptions.

AI/ML APPLICATIONS IN CNS DRUG DISCOVERY

CNS diseases are a group of neurological disorders that impose a significant economic and social impact. Development of new drugs for CNS diseases poses unique challenges compared to other diseases, including the complexity of brain anatomy and function, incomplete understanding of the biology of the complex nature of CNS diseases and the presence of BBB. In this section, we present an overview of AI/ML‐based approaches to meet challenges such as BBB permeability in CNS drug discovery (Figure 4).

Figure 4 AI/ML‐enabled improvements in the treatment of CNS diseases. DL is a subset of ML, which is a subset of AI and their applications address a wide range of challenges in CNS drug discovery and development. The application fields portrayed here are discussed in the Section 3. AI, artificial intelligence; CNS, central nervous system; DL, deep learning; ML, machine learning (Figure not reproduced here; see the original article.)

BBB permeability prediction

Despite significant progress in our understanding of CNS diseases, the development of novel therapies for CNS diseases faces some great challenges. In addition to the difficulties in CNS target identification, designing new molecules with the ability to penetrate the BBB is also a major obstacle. The role of the BBB is to protect the brain from variations in blood composition (e.g., hormones, amino acids, and potassium) and circulating pathogens. It consists of capillary endothelial cells that are lined by the basal lamina made from structural proteins (i.e., extracellular matrix proteins collagen and laminin), pericytes, astrocytic endfeet, and microglial cells. ³²⁷ This biologic membrane allows the uptake of water, glucose, and essential amino acids, the efflux of small molecules and nonessential amino acids from the brain to the blood and the passage of some molecules by passive diffusion. ³²⁸ While negligible penetration is desirable to minimize the brain side effects for peripheral drugs, high penetration is needed for CNS‐active drugs. To improve success rates in CNS drug discovery, the BBB permeability of drug candidates needs to be addressed early in the drug discovery process.

In recent years, AI‐based predictive models have been proposed to minimize the number of laborious, expensive, time‐consuming BBB permeability experiments that need be carried out in CNS drug discovery. For the construction of BBB permeability predictive models, researchers have employed various supervised learning approaches, such as SVM, ²²² , ³²⁹ , ³³⁰ , ³³¹ , ³³² , ³³³ recursive partitioning (RP), ³³⁴ , ³³⁵ Gaussian process, ³³⁶ DT, ³³⁷ KNN, ³³⁸ linear discriminant analysis, ³³⁹ consensus classifier, ³⁴⁰  and ANN. ³⁴¹ , ³⁴² , ³⁴³ All of these methods were developed to process physical and chemical features, which mainly include molecular weight, hydrophilicity (ClogP), lipophilicity (ClogD), topological polar surface area, acidic and basic atoms numbers, hydrogen bond donors and acceptors, water‐accessible volume, flexibility (rotatable bonds), van der Waals volume, and ionization potential.

The predictive capability of all the methods mentioned above is limited to passive diffusional uptake and predominantly relies on few molecular descriptors. However, many molecules, for example, glucose and insulin, pass BBB via complex mechanisms that involve specific drug‐transporter/drug‐receptor interactions. ³⁴⁴ , ³⁴⁵ Hence, such mechanisms are hard to be described by simple physicochemical features of compounds. Moreover, achieving therapeutic drug concentrations in CNS may be limited by membrane transporters such as the ATP‐binding cassette and efflux transporter P‐glycoprotein (P‐gp), ³⁴⁶ which mediates efflux of drugs from the BBB. Although the primary role of these efflux transporters is limiting the brain entry of neurotoxins, they also limit the entry of many therapeutics and may contribute to CNS pharmacoresistance. ³⁴⁷ , ³⁴⁸ Therefore, prediction methods need to both overcome the limitations of physicochemical features and address the multiple mechanisms associated with the drugs that pass the barrier and sustain in the brain. For this purpose, Yuan et al. ³³³ developed an SVM model by combining physicochemical properties and molecular fingerprints: The former is related to passive diffusion while the latter is associated with specific interactions, such as uptake, efflux, and protein binding. When compared to other SVM‐based BBB permeability predictors, the improved accuracy of their model shows that integration of the physicochemical properties and fingerprints can yield better predictions. Actually, all the AI/ML‐based models we have mentioned so far have been trained only on molecular properties disregarding the other types of information related to the efficacy of CNS drugs.

Clinical trials of many drug candidates generate a large amount of phenotypic data in CNS, but the relationship between the CNS side‐effects of drugs and their BBB permeation has not been adequately captured. To bridge the knowledge gap, Gao et al. ³⁴⁹ developed a BBB permeability prediction tool utilizing drug clinical phenotypes (drug side effects and drug indications). Although they explored the BBB permeability prediction from a new angle by accounting for passive diffusion as well as putative contributions of active transport and other complex mechanisms, the accuracy of their SVM method still needs to be improved. In fact, the features based on physics and chemistry are different; hence, the relation between drug side effects and therapeutic effects is more abstract and deeper. ³⁵⁰ For this reason, classical classification algorithms are not able to efficiently explore the relationship between data and results. On the contrary, DL architectures have the ability to extract useful information from complex data structures with abstract relationships. Therefore, Miao et al. ³⁵⁰ built a DL model to predict the BBB permeability of drugs based on clinical features and achieved better performance than the other existing methods.

AI/ML applications in drug discovery for neurological disorders

AI/ML applications in drug discovery for neurodevelopmental disorders

Schizophrenia is arguably the most puzzling of psychiatric disorders. ³⁵¹ As a neurodevelopmental disorder, ³⁵² schizophrenia shows a lifetime prevalence of 0.30%–0.66%, ³⁵³ generally beginning before age 25 years and persisting throughout life, making it one of the leading factors of global disease burden. ³⁵⁴ Despite more than a century of research, its complex pathophysiology remains unknown, ³⁵⁵ and currently, there is no effective drug for schizophrenia. Therefore, there is a need for alternative strategies to develop innovative drug treatments for schizophrenia. ³⁵⁶ In recent years, AI/ML has seen as a promising technology to inform schizophrenia diagnosis, ³⁵⁵ , ³⁵⁷ detecting heterogeneity, ³⁵⁸ , ³⁵⁹ , ³⁶⁰ subtyping, ³⁶¹ , ³⁶² and treatment.

In drug discovery studies for schizophrenia, researchers have utilized AI/ML methods with various purposes, including drug target identification, ³⁶³ , ³⁶⁴ developing QSAR models, ³⁶⁵ predicting monitoring dosing compliance, ³⁶⁶ predicting GPCRs targeting compounds, ³⁶⁴ and drug repositioning. ³⁶⁷ Specifically, schizophrenia target genes were identified based on publicly available microarray data sets using an SVM‐RFE (recursive feature elimination)‐based feature selection, where the genes initially ranked by an SVM classifier and the signature was then identified by discarding the genes that were not differentially expressed. To detect optimal biomarkers of presynaptic dopamine overactivity, which may cause schizophrenia, an SVM classifier was used. ³⁶³ SVM classifiers were also used to predict QSAR models of the GABA (gamma aminobutyric acid) uptake inhibitor drugs, which can be beneficial in the treatment of schizophrenia. ³⁶⁵ Moreover, SVM outperformed the other ML methods in predicting the repositioning drugs for schizophrenia when trained on drug expression profiles. ³⁶⁷ On the contrary, for schizophrenia subtyping, an unsupervised learning approach, multi‐view clustering, was employed by combining transcriptomic data with clinical phenotypes. ³⁶⁸ Setting a good example of the beneficiary of AI/ML in clinical drug trials, a novel AI platform AiCure ³⁶⁶ on mobile devices was used to assess the dosing compliance in Phase 2 clinical trial in schizophrenia patients. It, simply, confirms the medication ingestion visually by using facial recognition and computer vision.

One of the major obstacles in developing AI/ML methods for schizophrenia drug discovery is data availability. ³⁶⁹ Publicly available, large‐scale, well‐structured information on neural phenotypes, genomics, and clinical stages are greatly lacking, which arouses questions for the generalizability of AI/ML algorithms across different data sets without performance loss. However, the availability of such integrative databases can encourage the development of AI/ML‐based methods to investigate personalized therapies by solving the disease heterogeneity.

Another neurodevelopmental disorder is autism spectrum disorder (ASD), which is characterized by deficits in social communication and social interaction and the presence of restricted, repetitive patterns in behaviors or interests. ³⁷⁰ ML methods have been utilized in ASD research for improving the diagnosis ³⁷¹ and prognosis prediction. ³⁷¹ Also, there are few ML applications in drug discovery for ASD. For example, ML‐based cluster analysis (i.e., affinity propagation and k‐medoids) of clinical data (i.e., signs and biomarkers) exhibited a good performance in drug response prediction of ASD patients. ³⁷² Moreover, Bayesian ML models trained on HTS data revealed the potential repurposing of nicardipine or other dihydropyridine calcium channel antagonists for the treatment of Pitt Hopkins Syndrome, a rare genetic disorder that exhibits features of autistic spectrum disorders. ³⁷³ Recently, ML algorithms have been employed to predict the functional effects of variants in voltage‐gated sodium and calcium ion channels, which have been associated with ASD, schizophrenia and developmental encephalopathy. ³⁷⁴ Being trained on sequence‐ and structure‐based features, the ML model predicted the gain or loss of function effects of likely pathogenic missense variants in ion channels and the results were validated in exome‐wide data. On the contrary, the toxic compounds may trigger the recent increases in neurodevelopmental disorders among children. ³⁷⁵ To identify developmental neurotoxicants, researchers developed ML algorithms to predict the neurodevelopmental toxicity of compounds. ³⁷⁶ , ³⁷⁷

AI/ML applications in drug discovery for depression

AI/ML‐based methods have been utilized in psychiatric drug discovery, especially for pharmacological decision support. ³⁶⁷ , ³⁷⁸ , ³⁷⁹ In a depression study, researchers have developed a gradient boosting machine using the predictors identified by the elastic net to predict whether a patient will achieve symptomatic remission using an antidepressant, citalopram. ³⁸⁰ This model was also successfully applied to an escitalopram treatment group of an independent clinical trial. ³⁷⁸ In the next study of Chekroud et al., ³⁸¹ they clustered the symptoms using an unsupervised learning approach (hierarchical clustering) and predict the responsiveness of each cluster to the treatment of different antidepressant drugs using the same model in the previous study. To provide decision support for clinicians to select the best drugs for a given cluster of symptoms, a web‐based application was designed. This AI‐based service is prospectively tested in hospital settings and thereby serve as a promising model for direct research translation. ³⁸²

On the contrary, the model of Chekroud et al. ³⁸⁰ has some limitations. The model only predicts whether a patient responds to a specific antidepressant without measuring the degree of antidepressant response. Since it was designed for only one antidepressant, the model is not capable of selecting the most effective drugs among various antidepressant candidates for patients. ³⁸³ To address these limitations, Chang et al. ³⁸³ developed an Antidepressant Response Prediction Network (ARPNet) model based on an NN architecture. Through the literature‐based and data‐driven feature selection process, ARPNet predicts the degree of antidepressant response, whether the patient will reach clinical remission from depression, and a patient's response to a combination of one or more antidepressants.

Electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) data also have been employed in predicting drug responses to treatments of depression. Zhdanov et al. ³⁸⁴ used an SVM classifier to accurately predict the outcome of escitalopram treatment using patients' EEG data at the baseline and after the first 2 weeks of treatment. To identify a robust signature from resting‐state EEG that would predict response to antidepressants, Wu et al. ³⁸⁵ designed an end‐to‐end prediction algorithm with a latent space model. They applied their algorithm, Sparse EEG Latent SpacE Regression (SELSER), to data from an imaging‐coupled, placebo‐controlled antidepressant study and identified an EEG signature of patient's response to antidepressant treatment (i.e., sertraline). Ichikawa et al. ³⁸⁶ aimed to develop a melancholic depressive disorder biomarker to extract critically important functional connections (FCs) from fMRI data. By combining two ML algorithms (i.e., L1‐regularized sparse canonical correlation analysis and sparse logistic regression), they developed a classifier for melancholic depressive disorder and found out that antidepressants had a heterogeneous effect on the identified FCs of melancholic depressive disorder.

Although some of the recent AI/ML‐aided tools have been rapidly translated into the clinical trials, the AI/ML methods still are not used widely in clinical practice, while AI has been employed in psychiatric research over 20 years. ³⁸⁷ To close the gap between research and clinic, we need to improve the validity of diagnostic and prognostic labels, representability of the features, and generalizability of models. ³⁸⁸ As scientists continue to work to bridge the gap between research and clinic, it will be possible to provide efficient, personalized treatments based on a patient's unique characteristics. ³⁸⁹

AI/ML applications in drug discovery for Parkinson's disease

Parkinson's disease (PD) is the second most common age‐related neurodegenerative disorder, affecting over 1% of the population above the age of 60, increasing to 5% in individuals above 85 years of age. ³⁹⁰ PD is a prime example of a multifaceted disease, including a broad range of motor and non‐motor symptoms and possible contribution of genetic and environmental risk factors. ³⁹¹ Currently, there is no treatment to prevent the progressive depletion of dopaminergic neurons in the substantia nigra that underlies the movement control and cognitive loss, which is manifested with tremors and memory loss. ³⁹² , ³⁹³ Available drug treatments are based on the administration of levodopa (l‐dopa) and catechol‐O‐methyltransferase or monoamine oxidase B inhibitors, offering only symptomatic relief to the patients. ³⁹²

In PD research, previous AI applications have focused on diagnostic biomarker discovery in cerebrospinal fluid (CSF) and blood ³⁹⁴ , ³⁹⁵ , ³⁹⁶ , ³⁹⁷ and remote monitoring of treatment response by using electronic wearables. ³⁹⁸ , ³⁹⁹ , ⁴⁰⁰ , ⁴⁰¹ , ⁴⁰² On the contrary, recently, AI/ML has received little attention in PD drug discovery. Particularly, Shao et al. ⁴⁰³ initially built SVM models to quickly select the compounds containing indole–piperazine–pyrimidine scaffold among large chemical databases and subsequently identified novel compounds that simultaneously bind the two receptors—adenosine A2A receptor and dopamine D2 receptor—implicated in the PD pathophysiology. In another study, Sebastián‐Pérez ⁴⁰⁴ utilized several ML techniques to infer QSAR models for the identification of putative inhibitors of LRRK2 protein, a key genetic risk factor for familiar and sporadic PD. Moreover, AI‐based technologies have helped overcome the drug side effects in PD treatment. While l‐dopa has remained the cornerstone of PD therapy for reducing the symptoms associated with dopamine deficiency, almost half of PD patients treated with it eventually develop levodopa‐induced dyskinesia (LID), a side effect that causes abnormal involuntary movements. In a review paper, Johnston et al. ⁴⁰⁵ discussed the use of AI platforms to identify repurposing candidates for LID treatment and highlighted the potential of AI approaches by designing a drug repositioning case study. To identify novel repurposing candidates that may reduce LID, they utilized a literature mining approach based on an IBM Watson engine, where the semantic similarity and a “graph diffusion” algorithm were applied to score and rank each candidate drug.

Along with the identification of novel and repurposing candidates, AI/ML techniques have been applied to the development of in vitro and in vivo PD models for drug screening. Monzel et al. ⁴⁰⁶ created a human midbrain organoid model of PD as an in vitro toxicity assay and built an RF classifier to predict compounds' neurotoxic effect on organoids based on cellular features. To establish an efficient drug‐testing route, Hughes et al. ⁴⁰⁷ developed a zebrafish model of PD together with an AI/ML‐based method to classify movement disorders in this model using high‐resolution video captures. Encouraging results of all the studies discussed above, highlight the potential benefits of AI/ML applications for the discovery of efficient and multitargeting drugs against emerging targets in PD as well as the screening of the drug effects on PD models.

AI/ML applications in drug discovery for Alzheimer's disease

Increasing life expectancy has produced a dramatic rise in the prevalence, and thus impact, of aging‐related diseases. The most prevalent neurodegenerative disease in older adults is Alzheimer's disease (AD), characterized by insidious and progressive impairment of behavioral and cognitive functions, including memory. ⁴⁰⁸ The cause of AD is still unclear; however, generally accepted neuropathological hallmarks of AD include extracellular A‐beta plaques and intracellular neurofibrillary tangles, along with neuronal and synaptic loss and/or dysfunction. ⁴⁰⁹ Current drugs for AD target cholinergic and glutamatergic neurotransmission, thus improving symptoms, although they show limited benefits to most AD patients. ⁴¹⁰ Therefore, new treatments are urgently needed to prevent or delay disease onset, slow its progression, or improve patients' symptoms. ⁴¹¹ However, drug development for AD has been extraordinarily difficult, with a failure rate of over 99% and no new drug approved since 2003. ⁴¹¹ AD drug failures are likely due to the lack of sufficient target engagement and toxicity, while drug discovery efforts mainly challenged by an incomplete understanding of AD pathogenesis, multifactorial etiology, and complex pathophysiology.

In recent years, AI/ML‐based models have become popular in AD research, mostly utilizing for AD diagnosis and prognosis in dealing with electronic health records and images. ⁴¹² On the contrary, AI/ML techniques have not been widely employed in AD drug discovery. However, there have been a few studies that show the potential benefits of AI/ML applications for the discovery of AD drugs. ML approaches have assisted the target identification and characterization in AD, which is the initial phase of drug discovery. For example, Cordax ⁴¹³ (https://cordax.switchlab.org) is a novel structure‐based amyloid core sequence prediction method that implements ML to detect aggregation‐prone regions in proteins as well as to predict the structural topology, orientation and overall architecture of the resulting amyloid core. As an aggregation predictor, it uses structural information on amyloid cores currently available in the protein databank and translates structural compatibility and interaction energies into sequence aggregation propensity using logistic regression. Along with the characterization of amyloid fibrils, ML approaches have been utilized for identifying potential drug targets. HENA, ⁴¹⁴ a heterogeneous network‐based data set for AD, integrates distinct data types (i.e., PPI, gene coexpression, epistasis, genome‐wide association study, gene expression in different brain regions, and positive selection data) through GCN to predict AD‐associated genes.

Researchers have built ML models—SVM, ANN, and RF—to predict the inhibitory effect of compounds against AD‐related proteins—histone deacetylase (HDAC), ⁴¹⁵ acetylcholinesterase (AChE), ⁴¹⁶ and S100 calcium‐binding protein A9 (S100A9), ⁴¹⁷ respectively. Although these target‐specific models were successful for predicting the bioactive compounds, a high level of reliability is necessary for prioritizing compounds that are ultimately translated into assays. To generate hyper‐predictive ML models, Jamal et al. ⁴¹⁸ have included dynamic properties of compounds and protein–ligand interactions. Extracting the dynamic descriptors from molecular dynamics simulations of caspase‐8 ligand complexes to train ANN and RF models, they predicted the active compounds against caspase‐8, which plays a key role in causing AD. The major challenge in developing such predictive models of inhibitor activity is the lack of data on true‐negative compound–protein interactions. To address this challenge, Miyazaki et al. ⁴¹⁹ constructed a graph CNN model to explore compounds specifically targeting proteins without using the information on the true‐negative interaction and applied the model to identify inhibitors of BACE1 enzyme, a major target for AD.

Although these ML applications have advanced the discovery of single‐target inhibitors, the complex nature of AD requires the discovery of multitarget drugs to address the multiple pathways contributing to disease pathogenesis. Therefore, researchers have developed ML algorithms for predicting multitarget‐directed compounds against AD. Kleandrova et al. ⁴²⁰ designed seven molecules as triple target inhibitors of AD‐related proteins, namely GSK3B, HDAC1, and HDAC6 by combining perturbation theory and ML‐based on ANN. Using a new multitask QSAR model based on the linear discriminant analysis, Concu et al. ⁴²¹ predicted the inhibitors of the two isoforms of the monoamine oxidase (MAO) enzymes, MAO‐A and MAO‐B, which are involved in the pathology of AD, PD, and other neuropsychiatric disorders. As epigenetic therapeutics for AD, HDAC inhibitors have shown promise; however, nonspecificity and nonselectivity are the major problems of current HDAC inhibitors. Therefore, Gupta et al. ⁴²² combined VS and ML to classify the HDAC inhibitors and identified a novel compound that potentially inhibits all isoforms of class I and class IIb HDAC for AD therapy. In addition to these, Fang et al. ⁴²³ built 100 binary classifiers based on the naive Bayesian and RP algorithms to predict active small molecules against 25 key targets toward AD. Experimental validation of the predicted molecules yielded a compound that is a dual cholinesterase inhibitor and H3R antagonist. In their following study, ⁴²⁴ the system has been updated by assembling 204 binary classifiers towards 54 critical targets related to AD and the information of the classifiers was shared in a web server named AlzhCPI. Utilizing this classifier system, another group of researchers ⁴²⁵ has identified multiple targets of a traditional Chinese herbal medicine formula, Naodesheng, for application to AD. Natural products has continued to generate an increased interest as a mean of discovering novel bioactive compounds against AD. Grisoni et al. ⁴²⁶ proposed a VS protocol based on ML models to explore the bioactive synthetic mimetics of the natural product galantamine, which is the first natural product‐based AD drug approved by the FDA in 2001. ⁴²⁷ Using an ML‐based selection and target profiling program, they identified galantamine‐mimetic small molecules with multitarget activity on enzymes and receptors related to AD.

Besides the predictions of multitarget compounds based on their bioactivity against known drug targets in AD, Jamal et al. ⁴²⁸ predicted small molecules that show a high binding affinity for ML‐inferred possible therapeutic targets. Unlike previous studies that target known AD‐related proteins, they initially predicted the probable AD‐associated genes using ML classifiers that are trained on network, sequence and functional features. Then, they used a conventional VS tool to select the compounds that have high affinity for the majority of the predicted targets.

In addition to applications for identifying small molecules towards therapeutic targets for AD, ML techniques also have been utilized in drug repositioning efforts. For example, telmisartan has been associated with AD by a network‐based classification model. ³¹⁰

AI/ML approaches have also been applied to drug response studies to treat AD patients in a more precise, personalized way. Hampel et al. ⁴²⁹ has built an AI/ML‐based precision medicine framework for identifying the genomic biomarkers of response to AD therapy. Specifically, they studied blarcamesine (ANAVEX2‐73), a selective sigma‐1 receptor agonist, in a Phase 2a trial, where they obtained the patients' whole‐exome and transcriptome data and recorded the measures of safety, clinical features, pharmacokinetics, and efficacy. They analyzed the relationship between the patient data and efficacy outcome measures using unsupervised formal concept analysis, which ultimately identified the biomarkers of drug response. On the contrary, Lu et al. ⁴³⁰ evaluated the therapeutic effects of Dengzhan Shengmai formula, a traditional Chinese medicine, on AD patients by analyzing the diffusion tensor imaging data with ML. Their ML classifier revealed significant white‐matter network alterations after treatment.

AI/ML applications in anesthesia and pain treatment

The CNC drugs include general anesthetics and the analgesics, as well. In the past few years, we have witnessed the widespread use of autonomous and AI‐based recommender systems in therapeutic decision making in anesthesia and pain management. Especially, pharmacological robots have become an integral part of the anesthesia field, offering a personalized anesthetic drug dosage for maintaining patient homeostasis during general anesthesia and sedation. ⁴³¹ These robots use complex ML algorithms based on patient data (e.g., EEG monitor, blood pressure, heart rate, etc.) and pharmacokinetic features of drugs to provide the optimal drug dosage. The role of pharmacological robots and even more intelligent autonomous systems (i.e., cognitive robot, which can recognize crucial clinical state that requires human intervention) in the anesthesia field has been comprehensively overviewed by Cédrick et al. ⁴³² Besides the robotic systems, ML applications assisted the clinicians ⁴³³ to monitor the drug‐specific anesthetic states ⁴³⁴ , ⁴³⁵ , ⁴³⁶ and predict the adverse outcomes in anesthesia patients. ⁴³⁷ , ⁴³⁸ , ⁴³⁹

Similar to the anesthesia field, AI models have mainly utilized for clinical decision support in pain management. With the increasing amount of data collected by state‐of‐the‐art monitoring sensors and the Internet of Things, the AI‐assisted patient‐controlled analgesia has a great potential for personalized pain therapy. ⁴⁴⁰ The other clinical applications of AI systems in pain management include prediction of pain severity/modality and analgesic requirements, ⁴⁴¹ , ⁴⁴² , ⁴⁴³ individualized medicine decision support in analgesic treatment, ⁴⁴⁴ , ⁴⁴⁵ prediction of the effectiveness of the analgesics, ⁴⁴⁶ , ⁴⁴⁷ and prediction of medication overuse. ⁴⁴⁸ , ⁴⁴⁹ , ⁴⁵⁰ Besides the clinical applications, researchers have employed ML methods at the early stages of analgesic discovery, such as identifying novel genes and pathways associated with acute and chronic pain ⁴⁵¹ and predicting inhibitors of a drug target for pain (i.e., NaV1.7 sodium channel). ⁴⁵² To facilitate the prediction of novel multi‐target analgesics or drug combinations for pain treatment, researchers have established a comprehensive pain‐domain‐specific chemogenomics knowledgebase that includes the analgesics in current use, pain‐related targets with all available 3D structures, and the compounds reported for these target proteins. ⁴⁵³

CONCLUSIONS AND FUTURE DIRECTIONS

Given the complexity of neurological disorders, CNS drug development is still a long, expensive, inefficient, and challenging process with a low rate of new successful therapeutic discovery. To overcome the challenges of CNS drug discovery, researchers have utilized AI/ML‐based methods, which have played a promising role in all stages of drug discovery for a variety of diseases (Table 2). In general, AI/ML practices in pharmaceutical development have aroused great interest among researchers working in academia and industry. The number of start‐ups in this area has grown rapidly and reached 230 by June 2020. ⁴⁵⁴ Also, many pharmaceutical companies have invested in internal AI‐based research programs as well as in collaboration with AI start‐ups and academic institutions. ⁴⁵⁵ , ⁴⁵⁶ Recently, we have witnessed a massive collaborative effort by both academia and industry in response to the COVID‐19 outbreak. Labs and AI firms have shared their data and pipelines in open‐sourced platforms. For example, Google Deepmind has released the 3D structures of SARS‐CoV‐2 proteins that have been predicted by their AlphaFold system. Although AI‐enabled solutions have emerged as a crucial tool for transforming the process of therapeutic development, the use of AI technologies to improve CNS drug discovery is still at an early stage. Below, we discuss the limitations as well as the future directions to guide further advancement in this evolving field.

The main bottleneck in applying AI/ML into CNS drug discovery is the lack of high‐quality, well‐annotated data sets to train effective algorithms. The data collected in the public databases are generally generated by different biological assays, methods, or conditions, which are not comparable. Also, multiple data sets on the same subject may contradict each other. Therefore, filtering the raw inputs to obtain high‐quality data is a crucial step before performing specific AI/ML tasks.

The “black box” nature of most next‐generation AI architectures an additional challenge in CNS drug discovery. The lack of interpretability of AI/ML‐generated results limits their applications. While this is not the case for simpler ML models (i.e., XGBoost, TensorFlow, Lasso, Ridge, Elastic Net), for more advanced ML models (i.e., DNN) the internal workings remain a mystery. Hence, researchers cannot explain how the model arrives at the result and understand the underlying biological mechanisms. This also makes it more difficult to troubleshoot these models when they unexpectedly fail. Therefore, there is a critical need to develop methods for decoding the black boxes of DL.

Also, the amount of the available data directly affects the performance of AI/ML models, since successful training of algorithms relies on suitably large amounts of training data. This is a particular challenge in nominating new targets or drugs for many neurological conditions for which no treatment reverses the disease state. When there are not sufficient training examples for performing a drug discovery task, transfer learning technology, which learns from one task and applies it to the other task, can offer a solution. However, in the long term, the most promising solution to overcome data scarcity would be for the scientific community to share their data. Such large‐scale sharing of data would make significant improvement in the CNS drug discovery process, with advances in hardware that lead to faster machines such as quantum computers in the near future.

A particular limitation for the AI/ML applications in CNS drug discovery is the unknown pathophysiology for many nervous system disorders, which makes target identification very challenging. To explore the complex disease mechanisms and define the right biological targets, we need better AI/ML tools that can pull information out of the data sets generated across the different biological layers (e.g., transcriptomics, proteomics, and metabolomics). Here, capsule networks, ⁴⁵⁷ a next‐generation AI architecture where CNNs are encapsulated in an interconnected module, can provide a solution. As the first application of capsule networks to drug discovery, capsule networks showed excellent performance to predict the cardiotoxicity of compounds, which highlights their unique potential in drug discovery efforts. ²²⁶ Because of the modular representation of the CNNs, capsule networks can learn from heterogeneous data sets by preserving the hierarchical aspects of the data itself. Considering the highly modular nature of CNS data sets with specified layers of genes, proteins, metabolites, capsule networks can analyze the changes in the functional organization and interplay of these layers upon the diseases.

Another critical issue in the application of AI/ML models into CNS drug discovery is the integration of different data types, including genotypic data from patients, multiomics data from drug treatments, and chemical data from bioactivity and toxicity assays. Considering the availability of various databases that include biological, structural, and chemical information, how to integrate these data to generate AI/ML models becomes a critical question in CNS drug discovery applications. Multitask learning, learning of different tasks jointly, can be suitable for these types of applications. Multitask NNs are capable of integrating data from many distinct sources. For example, a multitask architecture can predict the effects of a drug and its BBB permeability at the same time by learning from multiomics data sets, physicochemical properties, HTS, and bioactivity assays.

In recent years, we have seen the emergence of novel neuroimaging techniques such as pharmacological functional magnetic resonance imaging (pharmacoMRI) and pharmacologically induced functional ultrasound (pharmaco‐fUS), which provide in vivo functional data of specific effects of drugs on the brain. Although pharmacoMRI continues to play a useful role in neuropharmacology studies as a well‐established technique, ⁴⁵⁸ , ⁴⁵⁹ , ⁴⁶⁰ , ⁴⁶¹ a variety of challenges (i.e., low sensitivity, the requirement for anesthesia, and blood oxygenation‐level dependent imaging) limit the preclinical use of it. A newer tool, pharmaco‐fUS enables brain activity imaging through the local monitoring of cerebral blood volume dynamics at an unprecedented spatiotemporal resolution without the bias of anesthesia. ⁴⁶² , ⁴⁶³ Recent studies demonstrated fUS imaging's potential to characterize dynamic profiles of CNS drugs, including a drug combination of donepezil plus mefloquine for AD ⁴⁶⁴ and atomoxetine for attention‐deficit/hyperactivity disorder. ⁴⁶⁵ Moreover, Rabut et al. ⁴⁶⁶ adapted ML to analyze the rich data content provided by fUS connectivity imaging. Their ML model identified the “fingerprint” of drug‐induced brain connectivity changes in awake mice for scopolamine, a major preclinical drug to model AD. As evident from the previous applications, AI/ML methods hold the promise of characterization of treatment effects from novel neuroimaging data sets and thereby improving our understanding of the mechanism of action of drugs in the brain. Getting drugs across the BBB is an essential step to developing successful therapies to treat CNS disorders. However, it is often overlooked that BBB is not only a physical barrier for drug delivery to the CNS but also a complex, dynamic interface that might be affected by diseases. CNS disorders may result in dysfunction of BBB, such as its disruption or dysfunctions related to BBB transporters. To date, AI/ML‐based predictive algorithms have assumed that BBB is a static entity by neglecting the effects of CNS pathologies on it. Therefore, a prediction model for BBB penetrance that trained on data from non‐CNS diseases may not work for a CNS disease. To develop better prediction models for BBB permeability, we need to take into account disease‐related changes in the barrier. This also provides many unique opportunities for developing disease‐specific AI/ML tools in CNS drug discovery.

It is important to highlight that CNS drug discovery has a nondeterministic nature, where the neurological targets involve different pathways and their biological consequences are not the sums of the single functions, most drugs have diverse activities through multiple biological targets, and drug response is dependent on a range of factors (i.e., patient's genetic profile and drug's membrane permeability). Moreover, physiologic events are highly context‐specific: A receptor interaction may take place in the liver but not in the brain. AI/ML systems often fail to pick up such context‐specific nonlinear relationships and many other unknown contributing factors. As a result of incomplete domain representation, partial predictability in CNS drug discovery is inevitable. For example, an AI/ML algorithm may predict drug targets that neuroscientists know will likely have significant side effects in the brain or generate unsynthesizable molecules. Here, we need the human refinement process and hypothesis‐driven approach ⁴⁶⁷ to address many of these challenges to achieve better performance. Knowledge acquisition from the human experts to the AI systems can help the AI/ML system learn and thereby guarantee the best scientific results. In consequence, this mixture of machine and mind ⁴⁶⁸ will improve decision making as an essential component of the CNS drug discovery process.

Although AI/ML algorithms have already revolutionized other fields, the adoption of them to drug discovery is still at an eraly stage. Initially, AI/ML algorithms have been developed and practically used for certain areas such as image recognition, gaming, and internet search. Inspired by the successful applications in other disciplines, scientists have applied AI/ML algorithms to pharmaceutical research. And yet, we do not have any AI/ML algorithm that is developed specifically for a drug discovery problem. But this means that there should be many opportunities to develop innovative and novel algorithms in the field of therapeutic discovery. In this way, AI/ML methods will play an increasingly important role in not just the field of general pharmaceutical research but also CNS drug discovery.

In conclusion, we extensively review the latest AI/ML‐assisted drug discovery applications for the therapy of CNS diseases. These applications have been overgrowing in the past couple of years, fueled by the unprecedented success of AI/ML‐based approaches in different fields of science and technology. We envision that in the future, AI/ML will play more and more critical roles in CNS drug discovery towards personalized medicine, especially in the following areas: (1) patient subtyping, (2) identification of key disease drivers, (3) prediction of cell type‐specific drug response, (4) autonomous design of novel drugs, and (5) disease‐specific BBB permeability testing. Today there are structural constraints in data and algorithms that are limiting the role of AI/ML. Nonetheless, in the long run, ongoing and emerging developments in AI/ML approaches to neuropharmacology will enable us to develop more effective drugs for CNS diseases.

Acknowledgements

ACKNOWLEDGMENTSThis study was supported in parts by grants from the National Institutes of Health (NIH)/National Institute on Aging (U01AG046170, RF1AG054014, RF1AG057440, R01AG057907, U01AG052411, R01AG062355, U01AG058635, R01AG068030) and the Mount Sinai seed fund to Bin Zhang.

References

1. Mohs RC , Greig NH . Drug discovery and development: role of basic biological research. Alzheimers Dement (N Y). 2017;3(4):651‐657.2925579110.1016/j.trci.2017.10.005PMC5725284 doi:10.1016/j.trci.2017.10.005 PubMed
2. Papadatos G , Gaulton A , Hersey A , Overington JP . Activity, assay and target data curation and quality in the ChEMBL database. J Comput Aided Mol Des. 2015;29(9):885‐896.2620139610.1007/s10822-015-9860-5PMC4607714 doi:10.1007/s10822-015-9860-5 PubMed
3. Wilson BJ , Nicholls SG . The Human Genome Project, and recent advances in personalized genomics. Risk Manag Healthc Policy. 2015;8:9‐20.2573393910.2147/RMHP.S58728PMC4337712 doi:10.2147/RMHP.S58728 PubMed
4. David L , Arús‐Pous J , Karlsson J , et al. Applications of deep‐learning in exploiting large‐scale and heterogeneous compound data in industrial pharmaceutical research. Front Pharmacol. 2019;10:1303.3174970510.3389/fphar.2019.01303PMC6848277 doi:10.3389/fphar.2019.01303 PubMed
5. Administration USFD. Artificial Intelligence and Machine Learning in Software as a Medical Device. https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device. Accessed October 26, 2020.
6. Todeschini R , Wiese M , Consonni V . Books—handbook of molecular descriptors. Angew Chem. 2001;40(10):1977.
7. Dong J , Cao D‐S , Miao H‐Y , et al. ChemDes: an integrated web‐based platform for molecular descriptor and fingerprint computation. J Cheminform. 2015;7(1):60.2666445810.1186/s13321-015-0109-zPMC4674923 doi:10.1186/s13321-015-0109-z PubMed
8. Cao DS , Zhou GH , Liu S , et al. Large‐scale prediction of human kinase‐inhibitor interactions using protein sequences and molecular topological structures. Anal Chim Acta. 2013;792:10‐18.2391096210.1016/j.aca.2013.07.003 doi:10.1016/j.aca.2013.07.003 PubMed
9. Yee SW , Lin L , Merski M , et al. Prediction and validation of enzyme and transporter off‐targets for metformin. J Pharmacokinet Pharmacodyn. 2015;42(5):463‐475.2633566110.1007/s10928-015-9436-yPMC4656030 doi:10.1007/s10928-015-9436-y PubMed
10. Muegge I , Mukherjee P . An overview of molecular fingerprint similarity search in virtual screening. Expert Opin Drug Discovery. 2016;11(2):137‐148.10.1517/17460441.2016.111707026558489 doi:10.1517/17460441.2016.1117070 PubMed
11. Cereto‐Massague A , Ojeda MJ , Valls C , Mulero M , Garcia‐Vallve S , Pujadas G . Molecular fingerprint similarity search in virtual screening. Methods (San Diego, Calif). 2015;71:58‐63.10.1016/j.ymeth.2014.08.00525132639 doi:10.1016/j.ymeth.2014.08.005 PubMed
12. Willett P . Similarity‐based virtual screening using 2D fingerprints. Drug Discovery Today. 2006;11(23‐24):1046‐1053.1712982210.1016/j.drudis.2006.10.005 doi:10.1016/j.drudis.2006.10.005 PubMed
13. Heikamp K , Bajorat J . Fingerprint design and engineering strategies: rationalizing and improving similarity search performance. Future Med Chem. 2012;4(15):1945‐1959.2308827510.4155/fmc.12.126 doi:10.4155/fmc.12.126 PubMed
14. Irwin JJ , Gaskins G , Sterling T , Mysinger MM , Keiser MJ . Predicted biological activity of purchasable chemical space. J Chem Inf Model. 2018;58(1):148‐164.2919397010.1021/acs.jcim.7b00316PMC5780839 doi:10.1021/acs.jcim.7b00316 PubMed
15. Axen SD , Huang XP , Caceres EL , Gendelev L , Roth BL , Keiser MJ . A simple representation of three‐dimensional molecular structure. J Med Chem. 2017;60(17):7393‐7409.2873133510.1021/acs.jmedchem.7b00696PMC6075869 doi:10.1021/acs.jmedchem.7b00696 PubMed
16. Geppert H , Vogt M , Bajorath J . Current trends in ligand‐based virtual screening: molecular representations, data mining methods, new application areas, and performance evaluation. J Chem Inf Model. 2010;50(2):205‐216.2008857510.1021/ci900419k doi:10.1021/ci900419k PubMed
17. Berenger F , Voet A , Lee XY , Zhang KY . A rotation‐translation invariant molecular descriptor of partial charges and its use in ligand‐based virtual screening. J Cheminform. 2014;6:23.2488717810.1186/1758-2946-6-23PMC4030740 doi:10.1186/1758-2946-6-23 PubMed
18. Roy K , Mitra I . Electrotopological state atom (E‐state) index in drug design, QSAR, property prediction and toxicity assessment. Curr Comput Aided Drug Des. 2012;8(2):135‐158.2249746910.2174/157340912800492366 doi:10.2174/157340912800492366 PubMed
19. Cao D‐S , Xu Q‐S , Liang Y‐Z , Chen X , Li H‐D . Prediction of aqueous solubility of druglike organic compounds using partial least squares, back‐propagation network and support vector machine. J Chemom. 2010;24(9):584‐595.
20. Viswanadhan VN , Rajesh H , Balaji VN . Atom type preferences, structural diversity, and property profiles of known drugs, leads, and nondrugs: a comparative assessment. ACS Comb Sci. 2011;13(3):327‐336.2148066910.1021/co2000168 doi:10.1021/co2000168 PubMed
21. Khan MT . Predictions of the ADMET properties of candidate drug molecules utilizing different QSAR/QSPR modelling approaches. Curr Drug Metab. 2010;11(4):285‐295.2045047710.2174/138920010791514306 doi:10.2174/138920010791514306 PubMed
22. Cheng F , Li W , Zhou Y , et al. Correction to “admetSAR: a comprehensive source and free tool for assessment of chemical ADMET properties”. J Chem Inf Model. 2019;59:4959.3166126210.1021/acs.jcim.9b00969 doi:10.1021/acs.jcim.9b00969 PubMed
23. Maltarollo VG , Gertrudes JC , Oliveira PR , Honorio KM . Applying machine learning techniques for ADME‐Tox prediction: a review. Expert Opin Drug Metab Toxicol. 2015;11(2):259‐271.2544052410.1517/17425255.2015.980814 doi:10.1517/17425255.2015.980814 PubMed
24. Chuang KV , Gunsalus LM , Keiser MJ . Learning molecular representations for medicinal chemistry. J Med Chem. 2020;63:8705‐8722.3236609810.1021/acs.jmedchem.0c00385 doi:10.1021/acs.jmedchem.0c00385 PubMed
25. Moffat JG , Vincent F , Lee JA , Eder J , Prunotto M . Opportunities and challenges in phenotypic drug discovery: an industry perspective. Nat Rev Drug Discovery. 2017;16(8):531‐543.2868576210.1038/nrd.2017.111 doi:10.1038/nrd.2017.111 PubMed
26. Van den Broeck WMM . Chapter 3. Drug targets, target identification, validation, and screening. In: Wermuth CG , Aldous D , Raboisson P , Rognan D , eds. The Practice of Medicinal Chemistry. 4th Ed. San Diego, CA: Academic Press; 2015:45‐70.
27. Gashaw I , Ellinghaus P , Sommer A , Asadullah K . What makes a good drug target? Drug Discovery Today. 2011;16(23–24):1037‐1043.2194586110.1016/j.drudis.2011.09.007 doi:10.1016/j.drudis.2011.09.007 PubMed
28. Gashaw I , Ellinghaus P , Sommer A , Asadullah K . What makes a good drug target? Drug Discovery Today. 2012;17(suppl):S24‐S30.2215564610.1016/j.drudis.2011.12.008 doi:10.1016/j.drudis.2011.12.008 PubMed
29. Lindsay MA . Target discovery. Nat Rev Drug Discovery. 2003;2(10):831‐838.1452638610.1038/nrd1202 doi:10.1038/nrd1202 PubMed
30. Sorlie T , Perou CM , Tibshirani R , et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA. 2001;98(19):10869‐10874.1155381510.1073/pnas.191367098PMC58566 doi:10.1073/pnas.191367098 PubMed
31. Shen R , Olshen AB , Ladanyi M . Integrative clustering of multiple genomic data types using a joint latent variable model with application to breast and lung cancer subtype analysis. Bioinformatics. 2009;25(22):2906‐2912.1975919710.1093/bioinformatics/btp543PMC2800366 doi:10.1093/bioinformatics/btp543 PubMed
32. Yuan Y , Savage RS , Markowetz F . Patient‐specific data fusion defines prognostic cancer subtypes. PLOS Comput Biol. 2011;7(10):e1002227.2202863610.1371/journal.pcbi.1002227PMC3197649 doi:10.1371/journal.pcbi.1002227 PubMed
33. Zhang L , Lv C , Jin Y , et al. Deep learning‐based multi‐omics data integration reveals two prognostic subtypes in high‐risk neuroblastoma. Front Genet. 2018;9:477.3040568910.3389/fgene.2018.00477PMC6201709 doi:10.3389/fgene.2018.00477 PubMed
34. Gao F , Wang W , Tan M , et al. DeepCC: a novel deep learning‐based framework for cancer molecular subtype classification. Oncogenesis. 2019;8(9):44.3142053310.1038/s41389-019-0157-8PMC6697729 doi:10.1038/s41389-019-0157-8 PubMed
35. Singh A , Shannon CP , Gautier B , et al. DIABLO: an integrative approach for identifying key molecular drivers from multi‐omics assays. Bioinformatics. 2019;35(17):3055‐3062.3065786610.1093/bioinformatics/bty1054PMC6735831 doi:10.1093/bioinformatics/bty1054 PubMed
36. Tenenhaus A , Tenenhaus M . Regularized generalized canonical correlation analysis. Psychometrika. 2011;76(2):257‐284.10.1007/s11336-017-9573-x28536930 doi:10.1007/s11336-017-9573-x PubMed
37. Argelaguet R , Velten B , Arnol D , et al. Multi‐omics factor analysis—a framework for unsupervised integration of multi‐omics data sets. Mol Syst Biol. 2018;14(6):e8124.2992556810.15252/msb.20178124PMC6010767 doi:10.15252/msb.20178124 PubMed
38. Lock EF , Hoadley KA , Marron JS , Nobel AB . Joint and individual variation explained (Jive) for integrated analysis of multiple data types. Ann Appl Stat. 2013;7(1):523‐542.2374515610.1214/12-AOAS597PMC3671601 doi:10.1214/12-AOAS597 PubMed
39. Wong WC , Kim D , Carter H , Diekhans M , Ryan MC , Karchin R . CHASM and SNVBox: toolkit for detecting biologically important single nucleotide mutations in cancer. Bioinformatics. 2011;27(15):2147‐2148.2168505310.1093/bioinformatics/btr357PMC3137226 doi:10.1093/bioinformatics/btr357 PubMed
40. Asif M , Martiniano H , Vicente AM , Couto FM . Identifying disease genes using machine learning and gene functional similarities, assessed through Gene Ontology. PLOS One. 2018;13(12):e0208626.3053219910.1371/journal.pone.0208626PMC6287949 doi:10.1371/journal.pone.0208626 PubMed
41. Mao Y , Chen H , Liang H , Meric‐Bernstam F , Mills GB , Chen K . CanDrA: cancer‐specific driver missense mutation annotation with optimized features. PLOS One. 2013;8(10):e77945.2420503910.1371/journal.pone.0077945PMC3813554 doi:10.1371/journal.pone.0077945 PubMed
42. Ghanat Bari M , Ung CY , Zhang C , Zhu S , Li H . Machine learning‐assisted network inference approach to identify a new class of genes that coordinate the functionality of cancer networks. Sci Rep. 2017;7(1):6993.2876556010.1038/s41598-017-07481-5PMC5539301 doi:10.1038/s41598-017-07481-5 PubMed
43. Botía JA , Guelfi S , Zhang D , et al. G2P: using machine learning to understand and predict genes causing rare neurological disorders. bioRxiv. 2018:288845. 10.1101/288845 doi:10.1101/288845
44. Han Y , Yang J , Qian X , et al. DriverML: a machine learning algorithm for identifying driver genes in cancer sequencing studies. Nucleic Acids Res. 2019;47(8):e45.3077359210.1093/nar/gkz096PMC6486576 doi:10.1093/nar/gkz096 PubMed
45. Chung IF , Chen CY , Su SC , et al. DriverDBv2: a database for human cancer driver gene research. Nucleic Acids Res. 2016;44(D1):D975‐D979.2663539110.1093/nar/gkv1314PMC4702919 doi:10.1093/nar/gkv1314 PubMed
46. Luo P , Ding Y , Lei X , Wu FX . deepDriver: predicting cancer driver genes based on somatic mutations using deep convolutional neural networks. Front Genet. 2019;10:13.3076118110.3389/fgene.2019.00013PMC6361806 doi:10.3389/fgene.2019.00013 PubMed
47. Peng J , Guan J , Shang X . Predicting Parkinson's disease genes based on Node2vec and autoencoder. Front Genet. 2019;10:226.3100131110.3389/fgene.2019.00226PMC6454041 doi:10.3389/fgene.2019.00226 PubMed
48. Collier O , Stoven V , Vert JP . LOTUS: a single‐ and multitask machine learning algorithm for the prediction of cancer driver genes. PLOS Comput Biol. 2019;15(9):e1007381.3156852810.1371/journal.pcbi.1007381PMC6786659 doi:10.1371/journal.pcbi.1007381 PubMed
49. Kim M , Tagkopoulos I . Data integration and predictive modeling methods for multi‐omics datasets. Mol Omics. 2018;14(1):8‐25.2972567310.1039/c7mo00051k doi:10.1039/c7mo00051k PubMed
50. Bravo A , Pinero J , Queralt‐Rosinach N , Rautschka M , Furlong LI . Extraction of relations between genes and diseases from text and large‐scale data analysis: implications for translational research. BMC Bioinformatics. 2015;16:55.2588673410.1186/s12859-015-0472-9PMC4466840 doi:10.1186/s12859-015-0472-9 PubMed
51. Bartel DP . MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004;116(2):281‐297.1474443810.1016/s0092-8674(04)00045-5 doi:10.1016/s0092-8674(04)00045-5 PubMed
52. Schmidt MF . miRNA targeting drugs: the next blockbusters? Methods Mol Biol (Clifton, NJ). 2017;1517:3‐22.10.1007/978-1-4939-6563-2_127924471 doi:10.1007/978-1-4939-6563-2_1 PubMed
53. Xu J , Li CX , Lv JY , et al. Prioritizing candidate disease miRNAs by topological features in the miRNA target‐dysregulated network: case study of prostate cancer. Mol Cancer Ther. 2011;10(10):1857‐1866.2176832910.1158/1535-7163.MCT-11-0055 doi:10.1158/1535-7163.MCT-11-0055 PubMed
54. Xuan P , Han K , Guo M , et al. Prediction of microRNAs associated with human diseases based on weighted k most similar neighbors. PLOS One. 2013;8(8):e70204.2395091210.1371/journal.pone.0070204PMC3738541 doi:10.1371/journal.pone.0070204 PubMed
55. Zeng X , Wang W , Deng G , Bing J , Zou Q . Prediction of potential disease‐associated microRNAs by using neural networks. Mol Ther Nucleic Acids. 2019;16:566‐575.3107793610.1016/j.omtn.2019.04.010PMC6510966 doi:10.1016/j.omtn.2019.04.010 PubMed
56. Zheng K , You ZH , Wang L , Zhou Y , Li LP , Li ZW . MLMDA: a machine learning approach to predict and validate microRNA‐disease associations by integrating of heterogenous information sources. J Transl Med. 2019;17(1):260.3139507210.1186/s12967-019-2009-xPMC6688360 doi:10.1186/s12967-019-2009-x PubMed
57. Black DL . Mechanisms of alternative pre‐messenger RNA splicing. Annu Rev Biochem. 2003;72:291‐336.1262633810.1146/annurev.biochem.72.121801.161720 doi:10.1146/annurev.biochem.72.121801.161720 PubMed
58. Barash Y , Vaquero‐Garcia J , González‐Vallinas J , et al. AVISPA: a web tool for the prediction and analysis of alternative splicing. Genome Biol. 2013;14(10):R114.2415675610.1186/gb-2013-14-10-r114PMC4014802 doi:10.1186/gb-2013-14-10-r114 PubMed
59. Leung MK , Xiong HY , Lee LJ , Frey BJ . Deep learning of the tissue‐regulated splicing code. Bioinformatics. 2014;30(12):i121‐i129.2493197510.1093/bioinformatics/btu277PMC4058935 doi:10.1093/bioinformatics/btu277 PubMed
60. Jha A , Gazzara MR , Barash Y . Integrative deep models for alternative splicing. Bioinformatics. 2017;33(14):i274‐i282.2888200010.1093/bioinformatics/btx268PMC5870723 doi:10.1093/bioinformatics/btx268 PubMed
61. Stanescu A , Tangirala K , Caragea D . Predicting alternatively spliced exons using semi‐supervised learning. Int J Data Min Bioinform. 2016;14(1):1‐21.
62. Xiong HY , Alipanahi B , Lee LJ , et al. RNA splicing. The human splicing code reveals new insights into the genetic determinants of disease. Science. 2015;347(6218):1254806.2552515910.1126/science.1254806PMC4362528 doi:10.1126/science.1254806 PubMed
63. Doncheva NT , Kacprowski T , Albrecht M . Recent approaches to the prioritization of candidate disease genes. Wiley Interdiscip Rev: Syst Biol Med. 2012;4(5):429‐442.2268953910.1002/wsbm.1177 doi:10.1002/wsbm.1177 PubMed
64. Jeon J , Nim S , Teyra J , et al. A systematic approach to identify novel cancer drug targets using machine learning, inhibitor design and high‐throughput screening. Genome Med. 2014;6:57.2516548910.1186/s13073-014-0057-7PMC4143549 doi:10.1186/s13073-014-0057-7 PubMed
65. Valentini G , Paccanaro A , Caniza H , Romero AE , Re M . An extensive analysis of disease‐gene associations using network integration and fast kernel‐based gene prioritization methods. Artif Intell Med. 2014;61(2):63‐78.2472603510.1016/j.artmed.2014.03.003PMC4070077 doi:10.1016/j.artmed.2014.03.003 PubMed
66. Ferrero E , Dunham I , Sanseau P . In silico prediction of novel therapeutic targets using gene‐disease association data. J Transl Med. 2017;15(1):182.2885137810.1186/s12967-017-1285-6PMC5576250 doi:10.1186/s12967-017-1285-6 PubMed
67. Kim J , So S , Lee HJ , Park JC , Kim JJ , Lee H . DigSee: disease gene search engine with evidence sentences (version cancer). Nucleic Acids Res. 2013;41(Web Server issue):W510‐W517.2376145210.1093/nar/gkt531PMC3692119 doi:10.1093/nar/gkt531 PubMed
68. Arabfard M , Ohadi M , Rezaei Tabar V , Delbari A , Kavousi K . Genome‐wide prediction and prioritization of human aging genes by data fusion: a machine learning approach. BMC Genomics. 2019;20(1):832.3170626810.1186/s12864-019-6140-0PMC6842548 doi:10.1186/s12864-019-6140-0 PubMed
69. Holbrook SR , Muskal SM , Kim S‐H . Predicting protein structural features with Artificial neural networks. Artif Intell Mol Biol. 1993:161‐194.
70. Cheng J , Tegge AN , Baldi P . Machine learning methods for protein structure prediction. IEEE Rev Biomed Eng. 2008;1:41‐49.2227489810.1109/RBME.2008.2008239 doi:10.1109/RBME.2008.2008239 PubMed
71. Kandathil SM , Greener JG , Jones DT . Recent developments in deep learning applied to protein structure prediction. Proteins. 2019;87(12):1179‐1189.3158978210.1002/prot.25824PMC6899861 doi:10.1002/prot.25824 PubMed
72. Tsuchiya Y , Tomii K . Neural networks for protein structure and function prediction and dynamic analysis. Biophys Rev. 2020;12:569‐573.3216661010.1007/s12551-020-00685-6PMC7242519 doi:10.1007/s12551-020-00685-6 PubMed
73. Torrisi M , Pollastri G , Le Q . Deep learning methods in protein structure prediction. Comput Struct Biotechnol J. 2020;18:1301‐1310.3261275310.1016/j.csbj.2019.12.011PMC7305407 doi:10.1016/j.csbj.2019.12.011 PubMed
74. AlQuraishi M . AlphaFold at CASP13. Bioinformatics. 2019;35(22):4862‐4865.3111637410.1093/bioinformatics/btz422PMC6907002 doi:10.1093/bioinformatics/btz422 PubMed
75. Wei G‐W . Protein structure prediction beyond AlphaFold. Nat Mach Intell. 2019;1(8):336‐337.10.1038/s42256-019-0086-4PMC1095638638515561 doi:10.1038/s42256-019-0086-4 PubMed
76. Powell A . Science unfolding in time: a situated protein folding landscape in retrospect. OSF Preprints. 2019. 10.31219/osf.io/4e8j7 doi:10.31219/osf.io/4e8j7
77. Zheng W , Li Y , Zhang C , Pearce R , Mortuza SM , Zhang Y . Deep‐learning contact‐map guided protein structure prediction in CASP13. Proteins. 2019;87(12):1149‐1164.3136514910.1002/prot.25792PMC6851476 doi:10.1002/prot.25792 PubMed
78. Wang S , Sun S , Li Z , Zhang R , Xu J . Accurate de novo prediction of protein contact map by ultra‐deep learning model. PLOS Comput Biol. 2017;13(1):e1005324.2805609010.1371/journal.pcbi.1005324PMC5249242 doi:10.1371/journal.pcbi.1005324 PubMed
79. Yang J , Anishchenko I , Park H , Peng Z , Ovchinnikov S , Baker D . Improved protein structure prediction using predicted interresidue orientations. Proc Natl Acad Sci. 2020;117(3):1496‐1503. 10.1073/pnas.1914677117 31896580PMC6983395 doi:10.1073/pnas.1914677117 PubMed
80. Wang T , Qiao Y , Ding W , Mao W , Zhou Y , Gong H . Improved fragment sampling for ab initio protein structure prediction using deep neural networks. Nat Mach Intell. 2019;1(8):347‐355.
81. Billings WM , Hedelius B , Millecam T , Wingate D , Corte DD . ProSPr: democratized implementation of alphafold protein distance prediction network. bioRxiv. 2019:830273. 10.1101/830273 doi:10.1101/830273
82. Ung PM , Rahman R , Schlessinger A . Redefining the protein kinase conformational space with machine learning. Cell Chem Biol. 2018;25(7):916‐924.2986127210.1016/j.chembiol.2018.05.002PMC6054563 doi:10.1016/j.chembiol.2018.05.002 PubMed
83. Matsumoto S , Ishida S , Araki M , Kato T , Terayama K , Okuno Y . Extraction of protein dynamics information hidden in Cryo‐EM map using deep learning. bioRxiv. 2020. 10.1101/2020.02.17.951863 doi:10.1101/2020.02.17.951863
84. Rost B , Radivojac P , Bromberg Y . Protein function in precision medicine: deep understanding with machine learning. FEBS Lett. 2016;590(15):2327‐2341.2742313610.1002/1873-3468.12307PMC5937700 doi:10.1002/1873-3468.12307 PubMed
85. Owens J . Determining druggability. Nat Rev Drug Discovery. 2007;6(3):187.
86. Nayal M , Honig B . On the nature of cavities on protein surfaces: application to the identification of drug‐binding sites. Proteins. 2006;63(4):892‐906.1647762210.1002/prot.20897 doi:10.1002/prot.20897 PubMed
87. Li Q , Lai L . Prediction of potential drug targets based on simple sequence properties. BMC Bioinformatics. 2007;8:353.1788383610.1186/1471-2105-8-353PMC2082046 doi:10.1186/1471-2105-8-353 PubMed
88. Bakheet TM , Doig AJ . Properties and identification of human protein drug targets. Bioinformatics. 2009;25(4):451‐457.1916430410.1093/bioinformatics/btp002 doi:10.1093/bioinformatics/btp002 PubMed
89. Costa PR , Acencio ML , Lemke N . A machine learning approach for genome‐wide prediction of morbid and druggable human genes based on systems‐level data. BMC Genomics. 2010;11(suppl 5):S9.10.1186/1471-2164-11-S5-S9PMC304580221210975 doi:10.1186/1471-2164-11-S5-S9 PubMed
90. Wang Q , Feng Y , Huang J , Wang T , Cheng G . A novel framework for the identification of drug target proteins: combining stacked auto‐encoders with a biased support vector machine. PLOS One. 2017;12(4):e0176486.2845357610.1371/journal.pone.0176486PMC5409512 doi:10.1371/journal.pone.0176486 PubMed
91. Yuan JH , Han SB , Richter S , Wade RC , Kokh DB . Druggability assessment in TRAPP using machine learning approaches. J Chem Inf Model. 2020;60(3):1685‐1699.3210547610.1021/acs.jcim.9b01185 doi:10.1021/acs.jcim.9b01185 PubMed
92. Dezső Z , Ceccarelli M . Machine learning prediction of oncology drug targets based on protein and network properties. BMC Bioinformatics. 2020;21(1):104.3217123810.1186/s12859-020-3442-9PMC7071582 doi:10.1186/s12859-020-3442-9 PubMed
93. Kandoi G , Acencio ML , Lemke N . Prediction of druggable proteins using machine learning and systems biology: a mini‐review. Front Physiol. 2015;6:366.2669690010.3389/fphys.2015.00366PMC4672042 doi:10.3389/fphys.2015.00366 PubMed
94. Reker D , Perna AM , Rodrigues T , et al. Revealing the macromolecular targets of complex natural products. Nat Chem. 2014;6(12):1072‐1078.2541188510.1038/nchem.2095 doi:10.1038/nchem.2095 PubMed
95. Reker D , Rodrigues T , Schneider P , Schneider G . Identifying the macromolecular targets of de novo‐designed chemical entities through self‐organizing map consensus. Proc Natl Acad Sci USA. 2014;111(11):4067‐4072.2459159510.1073/pnas.1320001111PMC3964060 doi:10.1073/pnas.1320001111 PubMed
96. Rodrigues T , Reker D , Kunze J , Schneider P , Schneider G . Revealing the macromolecular targets of fragment‐like natural products. Angew Chem Int Ed. 2015;54(36):10516‐10520.10.1002/anie.20150424126202212 doi:10.1002/anie.201504241 PubMed
97. Schneider G , Reker D , Chen T , Hauenstein K , Schneider P , Altmann KH . Deorphaning the macromolecular targets of the natural anticancer compound doliculide. Angew Chem Int Ed. 2016;55(40):12408‐12411.10.1002/anie.20160570727605391 doi:10.1002/anie.201605707 PubMed
98. NidhiGlick, M , Davies JW , Jenkins JL . Prediction of biological targets for compounds using multiple‐category Bayesian models trained on chemogenomics databases. J Chem Inf Model. 2006;46(3):1124‐1133.1671173210.1021/ci060003g doi:10.1021/ci060003g PubMed
99. Lee K , Lee M , Kim D . Utilizing random Forest QSAR models with optimized parameters for target identification and its application to target‐fishing server. BMC Bioinformatics. 2017;18(suppl 16):567.2929731510.1186/s12859-017-1960-xPMC5751401 doi:10.1186/s12859-017-1960-x PubMed
100. Madhukar NS , Khade PK , Huang L , et al. A Bayesian machine learning approach for drug target identification using diverse data types. Nat Commun. 2019;10(1):5221.3174508210.1038/s41467-019-12928-6PMC6863850 doi:10.1038/s41467-019-12928-6 PubMed
101. D'Souza S , Prema KV , Balaji S . Machine learning models for drug–target interactions: current knowledge and future directions. Drug Discovery Today. 2020;25(4):748‐756. 10.1016/j.drudis.2020.03.003 32171918 doi:10.1016/j.drudis.2020.03.003 PubMed
102. Anusuya S , Kesherwani M , Priya KV , et al. Drug‐target interactions: prediction methods and applications. Curr Protein Pept Sci. 2018;19(6):537‐561.2782935010.2174/1389203718666161108091609 doi:10.2174/1389203718666161108091609 PubMed
103. Yuan Q , Gao J , Wu D , Zhang S , Mamitsuka H , Zhu S . DrugE‐Rank: improving drug‐target interaction prediction of new candidate drugs or targets by ensemble learning to rank. Bioinformatics. 2016;32(12):i18‐i27.2730761510.1093/bioinformatics/btw244PMC4908328 doi:10.1093/bioinformatics/btw244 PubMed
104. Ezzat A , Wu M , Li X , Kwoh CK . Computational prediction of drug‐target interactions via ensemble learning. Methods Mol Biol (Clifton, NJ). 2019;1903:239‐254.10.1007/978-1-4939-8955-3_1430547446 doi:10.1007/978-1-4939-8955-3_14 PubMed
105. Ezzat A , Wu M , Li XL , Kwoh CK . Drug‐target interaction prediction using ensemble learning and dimensionality reduction. Methods (San Diego, Calif). 2017;129:81‐88.10.1016/j.ymeth.2017.05.01628549952 doi:10.1016/j.ymeth.2017.05.016 PubMed
106. Rayhan F , Ahmed S , Shatabda S , et al. iDTI‐ESBoost: identification of drug target interaction using evolutionary and structural features with boosting. Sci Rep. 2017;7(1):17731.2925528510.1038/s41598-017-18025-2PMC5735173 doi:10.1038/s41598-017-18025-2 PubMed
107. Pliakos K , Vens C . Drug‐target interaction prediction with tree‐ensemble learning and output space reconstruction. BMC Bioinformatics. 2020;21(1):49.3203353710.1186/s12859-020-3379-zPMC7006075 doi:10.1186/s12859-020-3379-z PubMed
108. Sharma A , Rani R . BE‐DTI': ensemble framework for drug target interaction prediction using dimensionality reduction and active learning. Comput Methods Programs Biomed. 2018;165:151‐162.3033707010.1016/j.cmpb.2018.08.011 doi:10.1016/j.cmpb.2018.08.011 PubMed
109. Wan F , Hong L , Xiao A , Jiang T , Zeng J . NeoDTI: neural integration of neighbor information from a heterogeneous network for discovering new drug‐target interactions. Bioinformatics. 2019;35(1):104‐111.3056154810.1093/bioinformatics/bty543 doi:10.1093/bioinformatics/bty543 PubMed
110. Monteiro NRC , Ribeiro B , Arrais J . Drug‐target interaction prediction: end‐to‐end deep learning approach. IEEE/ACM Trans Comput Biol Bioinform. 2020:1.10.1109/TCBB.2020.297733532142454 doi:10.1109/TCBB.2020.2977335 PubMed
111. Perlman L , Gottlieb A , Atias N , Ruppin E , Sharan R . Combining drug and gene similarity measures for drug‐target elucidation. J Comput Biol. 2011;18(2):133‐145.2131445310.1089/cmb.2010.0213 doi:10.1089/cmb.2010.0213 PubMed
112. Wang YC , Zhang CH , Deng NY , Wang Y . Kernel‐based data fusion improves the drug‐protein interaction prediction. Comput Biol Chem. 2011;35(6):353‐362.2209963210.1016/j.compbiolchem.2011.10.003 doi:10.1016/j.compbiolchem.2011.10.003 PubMed
113. Wang Y , Chen S , Deng N , Wang Y . Drug repositioning by kernel‐based integration of molecular structure, molecular activity, and phenotype data. PLOS One. 2013;8(11):e78518.2424431810.1371/journal.pone.0078518PMC3823875 doi:10.1371/journal.pone.0078518 PubMed
114. Yamanishi Y , Araki M , Gutteridge A , Honda W , Kanehisa M . Prediction of drug‐target interaction networks from the integration of chemical and genomic spaces. Bioinformatics. 2008;24(13):i232‐i240.1858671910.1093/bioinformatics/btn162PMC2718640 doi:10.1093/bioinformatics/btn162 PubMed
115. Nascimento AC , Prudencio RB , Costa IG . A multiple kernel learning algorithm for drug‐target interaction prediction. BMC Bioinformatics. 2016;17:46.2680121810.1186/s12859-016-0890-3PMC4722636 doi:10.1186/s12859-016-0890-3 PubMed
116. Ripphausen P , Nisius B , Peltason L , Bajorath J . Quo vadis, virtual screening? A comprehensive survey of prospective applications. J Med Chem. 2010;53(24):8461‐8467.2092925710.1021/jm101020z doi:10.1021/jm101020z PubMed
117. Clark DE . What has virtual screening ever done for drug discovery? Expert Opin Drug Discovery. 2008;3(8):841‐851.10.1517/17460441.3.8.84123484962 doi:10.1517/17460441.3.8.841 PubMed
118. Lionta E , Spyrou G , Vassilatis DK , Cournia Z . Structure‐based virtual screening for drug discovery: principles, applications and recent advances. Curr Top Med Chem. 2014;14(16):1923‐1938.2526279910.2174/1568026614666140929124445PMC4443793 doi:10.2174/1568026614666140929124445 PubMed
119. Ghosh S , Nie A , An J , Huang Z . Structure‐based virtual screening of chemical libraries for drug discovery. Curr Opin Chem Biol. 2006;10(3):194‐202.1667528610.1016/j.cbpa.2006.04.002 doi:10.1016/j.cbpa.2006.04.002 PubMed
120. Acharya C , Coop A , Polli JE , Mackerell AD Jr . Recent advances in ligand‐based drug design: relevance and utility of the conformationally sampled pharmacophore approach. Curr Comput Aided Drug Des. 2011;7(1):10‐22.2080718710.2174/157340911793743547PMC2975775 doi:10.2174/157340911793743547 PubMed
121. Maldonado AG , Doucet JP , Petitjean M , Fan B‐T . Molecular similarity and diversity in chemoinformatics: from theory to applications. Mol Divers. 2006;10(1):39‐79.1640452810.1007/s11030-006-8697-1 doi:10.1007/s11030-006-8697-1 PubMed
122. Keiser MJ , Irwin JJ , Shoichet BK . The chemical basis of pharmacology. Biochemistry. 2010;49(48):10267‐10276.2105865510.1021/bi101540gPMC2994275 doi:10.1021/bi101540g PubMed
123. Ballester PJ , Mitchell JB . A machine learning approach to predicting protein‐ligand binding affinity with applications to molecular docking. Bioinformatics. 2010;26(9):1169‐1175.2023694710.1093/bioinformatics/btq112PMC3524828 doi:10.1093/bioinformatics/btq112 PubMed
124. Zilian D , Sotriffer CA . SFCscore(RF): a random forest‐based scoring function for improved affinity prediction of protein‐ligand complexes. J Chem Inf Model. 2013;53(8):1923‐1933.2370579510.1021/ci400120b doi:10.1021/ci400120b PubMed
125. Liu Q , Kwoh CK , Li J . Binding affinity prediction for protein‐ligand complexes based on beta contacts and B factor. J Chem Inf Model. 2013;53(11):3076‐3085.2419169210.1021/ci400450h doi:10.1021/ci400450h PubMed
126. Li H , Leung KS , Ballester PJ , Wong MH . istar: a web platform for large‐scale protein‐ligand docking. PLOS One. 2014;9(1):e85678.2447504910.1371/journal.pone.0085678PMC3901662 doi:10.1371/journal.pone.0085678 PubMed
127. Li GB , Yang LL , Wang WJ , Li LL , Yang SY . ID‐Score: a new empirical scoring function based on a comprehensive set of descriptors related to protein‐ligand interactions. J Chem Inf Model. 2013;53(3):592‐600.2339407210.1021/ci300493w doi:10.1021/ci300493w PubMed
128. Ballester PJ . Machine learning scoring functions based on random forest and support vector regression. Paper presented at: IAPR International Conference on Pattern Recognition in Bioinformatics. Berlin, Heidelberg: Springer; 2012:14‐25. 10.1007/978-3-642-34123-6_2 doi:10.1007/978-3-642-34123-6_2
129. Durrant JD , McCammon JA . NNScore: a neural‐network‐based scoring function for the characterization of protein‐ligand complexes. J Chem Inf Model. 2010;50(10):1865‐1871.2084595410.1021/ci100244vPMC2964041 doi:10.1021/ci100244v PubMed
130. Ouyang X , Handoko SD , Kwoh CK . CScore: a simple yet effective scoring function for protein‐ligand binding affinity prediction using modified CMAC learning architecture. J Bioinform Comput Biol. 2011;9(suppl 1):1‐14.10.1142/s021972001100577x22144250 doi:10.1142/s021972001100577x PubMed
131. Cang Z , Wei GW . TopologyNet: topology based deep convolutional and multi‐task neural networks for biomolecular property predictions. PLOS Comput Biol. 2017;13(7):e1005690.2874996910.1371/journal.pcbi.1005690PMC5549771 doi:10.1371/journal.pcbi.1005690 PubMed
132. Ragoza M , Hochuli J , Idrobo E , Sunseri J , Koes DR . Protein‐ligand scoring with convolutional neural networks. J Chem Inf Model. 2017;57(4):942‐957.2836858710.1021/acs.jcim.6b00740PMC5479431 doi:10.1021/acs.jcim.6b00740 PubMed
133. Stepniewska‐Dziubinska MM , Zielenkiewicz P , Siedlecki P . Development and evaluation of a deep learning model for protein‐ligand binding affinity prediction. Bioinformatics. 2018;34(21):3666‐3674.2975735310.1093/bioinformatics/bty374PMC6198856 doi:10.1093/bioinformatics/bty374 PubMed
134. Ashtawy HM , Mahapatra NR . BgN‐Score and BsN‐Score: bagging and boosting based ensemble neural networks scoring functions for accurate binding affinity prediction of protein‐ligand complexes. BMC Bioinformatics. 2015;16(suppl 4):S8.10.1186/1471-2105-16-S4-S8PMC434762225734685 doi:10.1186/1471-2105-16-S4-S8 PubMed
135. Ain QU , Aleksandrova A , Roessler FD , Ballester PJ . Machine‐learning scoring functions to improve structure‐based binding affinity prediction and virtual screening. Wiley Interdiscip Rev Comput Mol Sci. 2015;5(6):405‐424.2711029210.1002/wcms.1225PMC4832270 doi:10.1002/wcms.1225 PubMed
136. Yang X , Wang Y , Byrne R , Schneider G , Yang S . Concepts of artificial intelligence for computer‐assisted drug discovery. Chem Rev. 2019;119(18):10520‐10594.3129497210.1021/acs.chemrev.8b00728 doi:10.1021/acs.chemrev.8b00728 PubMed
137. Wang D , Cui C , Ding X , et al. Improving the virtual screening ability of target‐specific scoring functions using deep learning methods. Front Pharmacol. 2019;10:924.3150742010.3389/fphar.2019.00924PMC6713720 doi:10.3389/fphar.2019.00924 PubMed
138. Li L , Khanna M , Jo I , et al. Target‐specific support vector machine scoring in structure‐based virtual screening: computational validation, in vitro testing in kinases, and effects on lung cancer cell proliferation. J Chem Inf Model. 2011;51(4):755‐759.2143854810.1021/ci100490wPMC3092157 doi:10.1021/ci100490w PubMed
139. Xu D , Li L , Zhou D , Liu D , Hudmon A , Meroueh SO . Structure‐based target‐specific screening leads to small‐molecule CaMKII inhibitors. ChemMedChem. 2017;12(9):660‐677.2837119110.1002/cmdc.201600636PMC5554713 doi:10.1002/cmdc.201600636 PubMed
140. Sun H , Pan P , Tian S , et al. Constructing and validating high‐performance MIEC‐SVM models in virtual screening for kinases: a better way for actives discovery. Sci Rep. 2016;6:24817.2710254910.1038/srep24817PMC4840416 doi:10.1038/srep24817 PubMed
141. Berishvili VP , Voronkov AE , Radchenko EV , Palyulin VA . Machine learning classification models to improve the docking‐based screening: a case of PI3K‐tankyrase inhibitors. Mol Inf. 2018;37(11):e1800030.10.1002/minf.20180003029901257 doi:10.1002/minf.201800030 PubMed
142. Yan Y , Wang W , Sun Z , Zhang JZH , Ji C . Protein‐ligand empirical interaction components for virtual screening. J Chem Inf Model. 2017;57(8):1793‐1806.2867848410.1021/acs.jcim.7b00017 doi:10.1021/acs.jcim.7b00017 PubMed
143. Cherkasov A , Hilpert K , Jenssen H , et al. Use of artificial intelligence in the design of small peptide antibiotics effective against a broad spectrum of highly antibiotic‐resistant superbugs. ACS Chem Biol. 2009;4(1):65‐74.1905542510.1021/cb800240j doi:10.1021/cb800240j PubMed
144. Kinnings SL , Liu N , Tonge PJ , Jackson RM , Xie L , Bourne PE . A machine learning‐based method to improve docking scoring functions and its application to drug repurposing. J Chem Inf Model. 2011;51(2):408‐419.2129117410.1021/ci100369fPMC3076728 doi:10.1021/ci100369f PubMed
145. Pereira JC , Caffarena ER , Dos Santos CN . Boosting docking‐based virtual screening with deep learning. J Chem Inf Model. 2016;56(12):2495‐2506.2802440510.1021/acs.jcim.6b00355 doi:10.1021/acs.jcim.6b00355 PubMed
146. Leong MK , Syu RG , Ding YL , Weng CF . Prediction of N‐methyl‐d‐aspartate receptor GluN1‐ligand binding affinity by a novel SVM‐Pose/SVM‐Score combinatorial ensemble docking scheme. Sci Rep. 2017;7:40053.2805913310.1038/srep40053PMC5216401 doi:10.1038/srep40053 PubMed
147. Li H , Leung KS , Wong MH , Ballester PJ . Improving Autodock Vina using Random Forest: the growing accuracy of binding affinity prediction by the effective exploitation of larger data sets. Mol Inf. 2015;34(2–3):115‐126.10.1002/minf.20140013227490034 doi:10.1002/minf.201400132 PubMed
148. Arciniega M , Lange OF . Improvement of virtual screening results by docking data feature analysis. J Chem Inf Model. 2014;54(5):1401‐1411.2479693610.1021/ci500028u doi:10.1021/ci500028u PubMed
149. Waszkowycz B . Towards improving compound selection in structure‐based virtual screening. Drug Discovery Today. 2008;13(5–6):219‐226.1834279710.1016/j.drudis.2007.12.002 doi:10.1016/j.drudis.2007.12.002 PubMed
150. Johnson MA , Maggiora GM . Concepts and Applications of Molecular Similarity. Wiley; 1990.
151. Carpenter KA , Cohen DS , Jarrell JT , Huang X . Deep learning and virtual drug screening. Future Med Chem. 2018;10:2557‐2567.3028899710.4155/fmc-2018-0314PMC6563286 doi:10.4155/fmc-2018-0314 PubMed
152. Melville JL , Burke EK , Hirst JD . Machine learning in virtual screening. Comb Chem High Throughput Screen. 2009;12(4):332‐343.1944206310.2174/138620709788167980 doi:10.2174/138620709788167980 PubMed
153. Stokes JM , Yang K , Swanson K , et al. A deep learning approach to antibiotic discovery. Cell. 2020;180(4):688‐702.3208434010.1016/j.cell.2020.01.021PMC8349178 doi:10.1016/j.cell.2020.01.021 PubMed
154. Wu Z , Ramsundar B , Feinberg EN , et al. MoleculeNet: a benchmark for molecular machine learning. Chem Sci. 2018;9(2):513‐530.2962911810.1039/c7sc02664aPMC5868307 doi:10.1039/c7sc02664a PubMed
155. Feinberg EN , Joshi E , Pande VS , Cheng AC . Improvement in ADMET prediction with multitask deep featurization. J Med Chem. 2020;63(16):8835‐8848.3228682410.1021/acs.jmedchem.9b02187 doi:10.1021/acs.jmedchem.9b02187 PubMed
156. Wang T , Wu MB , Lin JP , Yang LR . Quantitative structure‐activity relationship: promising advances in drug discovery platforms. Expert Opin Drug Discovery. 2015;10(12):1283‐1300.10.1517/17460441.2015.108300626358617 doi:10.1517/17460441.2015.1083006 PubMed
157. Kumar R , Chaudhary K , Singh Chauhan J , et al. An in silico platform for predicting, screening and designing of antihypertensive peptides. Sci Rep. 2015;5:12512.2621311510.1038/srep12512PMC4515604 doi:10.1038/srep12512 PubMed
158. Briard JG , Fernandez M , de Luna P , Woo TK , Ben RN . QSAR Accelerated discovery of potent ice recrystallization inhibitors. Sci Rep. 2016;6:26403.2721658510.1038/srep26403PMC4877635 doi:10.1038/srep26403 PubMed
159. Svetnik V , Liaw A , Tong C , Culberson JC , Sheridan RP , Feuston BP . Random forest: a classification and regression tool for compound classification and QSAR modeling. J Chem Inf Comput Sci. 2003;43(6):1947‐1958.1463244510.1021/ci034160g doi:10.1021/ci034160g PubMed
160. Zakharov AV , Varlamova EV , Lagunin AA , et al. QSAR modeling and prediction of drug‐drug interactions. Mol Pharmaceutics. 2016;13(2):545‐556.10.1021/acs.molpharmaceut.5b0076226669717 doi:10.1021/acs.molpharmaceut.5b00762 PubMed
161. Fang X , Bagui S , Bagui S . Improving virtual screening predictive accuracy of Human kallikrein 5 inhibitors using machine learning models. Comput Biol Chem. 2017;69:110‐119.2860176110.1016/j.compbiolchem.2017.05.007 doi:10.1016/j.compbiolchem.2017.05.007 PubMed
162. Chen JJF , Visco DP Jr . Developing an in silico pipeline for faster drug candidate discovery: virtual high throughput screening with the signature molecular descriptor using support vector machine models. Chem Eng Sci. 2017;159:31‐42.
163. Chen JJF , Visco DP Jr . Identifying novel factor XIIa inhibitors with PCA‐GA‐SVM developed vHTS models. Eur J Med Chem. 2017;140:31‐41.2891809810.1016/j.ejmech.2017.08.056 doi:10.1016/j.ejmech.2017.08.056 PubMed
164. Xia X , Maliski EG , Gallant P , Rogers D . Classification of kinase inhibitors using a Bayesian model. J Med Chem. 2004;47(18):4463‐4470.1531745810.1021/jm0303195 doi:10.1021/jm0303195 PubMed
165. Bender A , Mussa HY , Glen RC . Screening for dihydrofolate reductase inhibitors using MOLPRINT 2D, a fast fragment‐based method employing the Naive Bayesian classifier: limitations of the descriptor and the importance of balanced chemistry in training and test sets. J Biomol Screen. 2005;10(7):658‐666.1617005110.1177/1087057105281048 doi:10.1177/1087057105281048 PubMed
166. Prathipati P , Ma NL , Keller TH . Global Bayesian models for the prioritization of antitubercular agents. J Chem Inf Model. 2008;48(12):2362‐2370.1905351810.1021/ci800143n doi:10.1021/ci800143n PubMed
167. Ekins S , Reynolds RC , Kim H , et al. Bayesian models leveraging bioactivity and cytotoxicity information for drug discovery. Chem Biol. 2013;20(3):370‐378.2352179510.1016/j.chembiol.2013.01.011PMC3607962 doi:10.1016/j.chembiol.2013.01.011 PubMed
168. Vijayan RS , Bera I , Prabu M , Saha S , Ghoshal N . Combinatorial library enumeration and lead hopping using comparative interaction fingerprint analysis and classical 2D QSAR methods for seeking novel GABA(A) alpha(3) modulators. J Chem Inf Model. 2009;49(11):2498‐2511.1989142110.1021/ci900309s doi:10.1021/ci900309s PubMed
169. Liu L , Lu J , Lu Y , et al. Novel Bayesian classification models for predicting compounds blocking hERG potassium channels. Acta Pharmacol Sin. 2014;35(8):1093‐1102.2497615410.1038/aps.2014.35PMC4125710 doi:10.1038/aps.2014.35 PubMed
170. Singh N , Chaudhury S , Liu R , AbdulHameed MDM , Tawa G , Wallqvist A . QSAR classification model for antibacterial compounds and its use in virtual screening. J Chem Inf Model. 2012;52(10):2559‐2569.2301354610.1021/ci300336v doi:10.1021/ci300336v PubMed
171. Renault N , Laurent X , Farce A , et al. Virtual screening of CB(2) receptor agonists from bayesian network and high‐throughput docking: structural insights into agonist‐modulated GPCR features. Chem Biol Drug Des. 2013;81(4):442‐454.2321706010.1111/cbdd.12095 doi:10.1111/cbdd.12095 PubMed
172. AbdulHameed MD , Ippolito DL , Wallqvist A . Predicting rat and human pregnane X receptor activators using Bayesian classification models. Chem Res Toxicol. 2016;29(10):1729‐1740.2760367510.1021/acs.chemrestox.6b00227 doi:10.1021/acs.chemrestox.6b00227 PubMed
173. Shi H , Tian S , Li Y , et al. Absorption, distribution, metabolism, excretion, and toxicity evaluation in drug discovery. 14. Prediction of human pregnane X receptor activators by using naive bayesian classification technique. Chem Res Toxicol. 2015;28(1):116‐125.2549554210.1021/tx500389q doi:10.1021/tx500389q PubMed
174. Murcia‐Soler M , Pérez‐Giménez F , García‐March FJ , et al. Artificial neural networks and linear discriminant analysis: a valuable combination in the selection of new antibacterial compounds. J Chem Inf Comput Sci. 2004;44(3):1031‐1041.1515477210.1021/ci030340e doi:10.1021/ci030340e PubMed
175. Douali L , Villemin D , Cherqaoui D . Neural networks: accurate nonlinear QSAR model for HEPT derivatives. J Chem Inf Comput Sci. 2003;43(4):1200‐1207.1287091210.1021/ci034047q doi:10.1021/ci034047q PubMed
176. Sabet R , Fassihi A , Hemmateenejad B , Saghaei L , Miri R , Gholami M . Computer‐aided design of novel antibacterial 3‐hydroxypyridine‐4‐ones: application of QSAR methods based on the MOLMAP approach. J Comput Aided Mol Des. 2012;26(3):349‐361.2245685910.1007/s10822-012-9561-2 doi:10.1007/s10822-012-9561-2 PubMed
177. Fjell CD , Jenssen H , Hilpert K , et al. Identification of novel antibacterial peptides by chemoinformatics and machine learning. J Med Chem. 2009;52(7):2006‐2015.1929659810.1021/jm8015365 doi:10.1021/jm8015365 PubMed
178. Torrent M , Andreu D , Nogues VM , Boix E . Connecting peptide physicochemical and antimicrobial properties by a rational prediction model. PLOS One. 2011;6(2):e16968.2134739210.1371/journal.pone.0016968PMC3036733 doi:10.1371/journal.pone.0016968 PubMed
179. Sardari S , Kohanzad H , Ghavami G . Artificial neural network modeling of antimycobacterial chemical space to introduce efficient descriptors employed for drug design. Chemometr Intell Lab. 2014;130:151‐158.
180. Khatri N , Lather V , Madan A . Diverse classification models for anti‐hepatitis C virus activity of thiourea derivatives. Chemometr Intell Lab. 2015;140:13‐21.
181. Hu L , Chen G , Chau RM . A neural networks‐based drug discovery approach and its application for designing aldose reductase inhibitors. J Mol Graph Model. 2006;24(4):244‐253.1622691110.1016/j.jmgm.2005.09.002 doi:10.1016/j.jmgm.2005.09.002 PubMed
182. Patra JC , Chua BH . Artificial neural network‐based drug design for diabetes mellitus using flavonoids. J Comput Chem. 2011;32(4):555‐567.2080626210.1002/jcc.21641 doi:10.1002/jcc.21641 PubMed
183. Myint KZ , Wang L , Tong Q , Xie XQ . Molecular fingerprint‐based artificial neural networks QSAR for ligand biological activity predictions. Mol Pharmaceutics. 2012;9(10):2912‐2923.10.1021/mp300237zPMC346224422937990 doi:10.1021/mp300237z PubMed
184. Geanes AR , Cho HP , Nance KD , et al. Ligand‐based virtual screen for the discovery of novel M5 inhibitor chemotypes. Bioorg Med Chem Lett. 2016;26(18):4487‐4491.2750367810.1016/j.bmcl.2016.07.071PMC4996684 doi:10.1016/j.bmcl.2016.07.071 PubMed
185. Ma J , Sheridan RP , Liaw A , Dahl GE , Svetnik V . Deep neural nets as a method for quantitative structure‐activity relationships. J Chem Inf Model. 2015;55(2):263‐274.2563532410.1021/ci500747n doi:10.1021/ci500747n PubMed
186. Martin EJ , Polyakov VR , Tian L , Perez RC . Profile‐QSAR 2.0: kinase virtual screening accuracy comparable to four‐concentration IC50s for realistically novel compounds. J Chem Inf Model. 2017;57(8):2077‐2088.2865143310.1021/acs.jcim.7b00166 doi:10.1021/acs.jcim.7b00166 PubMed
187. Shamsara J . A random forest model to predict the activity of a large set of soluble epoxide hydrolase inhibitors solely based on a set of simple fragmental descriptors. Comb Chem High Throughput Screen. 2019;22:555‐569.3162221610.2174/1386207322666191016110232 doi:10.2174/1386207322666191016110232 PubMed
188. Simeon S , Jongkon N . Construction of quantitative structure activity relationship (QSAR) models to predict potency of structurally diverse Janus kinase 2 inhibitors. Molecules (Basel, Switzerland). 2019;24(23):4393.10.3390/molecules24234393PMC693064031805692 doi:10.3390/molecules24234393 PubMed
189. Marchese Robinson RL , Palczewska A , Palczewski J , Kidley N . Comparison of the predictive performance and interpretability of Random Forest and Linear Models on Benchmark Data Sets. J Chem Inf Model. 2017;57(8):1773‐1792.2871520910.1021/acs.jcim.6b00753 doi:10.1021/acs.jcim.6b00753 PubMed
190. Speck‐Planche A , Kleandrova VV , Cordeiro MN . New insights toward the discovery of antibacterial agents: multi‐tasking QSBER model for the simultaneous prediction of anti‐tuberculosis activity and toxicological profiles of drugs. Eur J Pharm Sci. 2013;48(4–5):812‐818.2337621110.1016/j.ejps.2013.01.011 doi:10.1016/j.ejps.2013.01.011 PubMed
191. Tenorio‐Borroto E , Peñuelas Rivas CG , Vásquez Chagoyán JC , et al. ANN multiplexing model of drugs effect on macrophages; theoretical and flow cytometry study on the cytotoxicity of the anti‐microbial drug G1 in spleen. Bioorg Med Chem. 2012;20(20):6181‐6194.2298191710.1016/j.bmc.2012.07.020 doi:10.1016/j.bmc.2012.07.020 PubMed
192. Tenorio‐Borroto E , Peñuelas‐Rivas CG , Vásquez‐Chagoyán JC , et al. Model for high‐throughput screening of drug immunotoxicity—study of the anti‐microbial G1 over peritoneal macrophages using flow cytometry. Eur J Med Chem. 2014;72:206‐220.2444528010.1016/j.ejmech.2013.08.035 doi:10.1016/j.ejmech.2013.08.035 PubMed
193. Speck‐Planche A , Cordeiro MN . Simultaneous modeling of antimycobacterial activities and ADMET profiles: a chemoinformatic approach to medicinal chemistry. Curr Top Med Chem. 2013;13(14):1656‐1665.2388905210.2174/15680266113139990116 doi:10.2174/15680266113139990116 PubMed
194. Speck‐Planche A , Cordeiro MN . Simultaneous virtual prediction of anti‐Escherichia coli activities and ADMET profiles: a chemoinformatic complementary approach for high‐throughput screening. ACS Comb Sci. 2014;16(2):78‐84.2438395810.1021/co400115s doi:10.1021/co400115s PubMed
195. Kleandrova VV , Ruso JM , Speck‐Planche A , Dias Soeiro Cordeiro MN . Enabling the discovery and virtual screening of potent and safe antimicrobial peptides. Simultaneous prediction of antibacterial activity and cytotoxicity. ACS Comb Sci. 2016;18(8):490‐498.2728073510.1021/acscombsci.6b00063 doi:10.1021/acscombsci.6b00063 PubMed
196. Speck‐Planche A , Dias Soeiro Cordeiro MN . Speeding up early drug discovery in antiviral research: a fragment‐based in silico approach for the design of virtual anti‐hepatitis C leads. ACS Comb Sci. 2017;19(8):501‐512.2843709110.1021/acscombsci.7b00039 doi:10.1021/acscombsci.7b00039 PubMed
197. Vina D , Uriarte E , Orallo F , Gonzalez‐Diaz H . Alignment‐free prediction of a drug‐target complex network based on parameters of drug connectivity and protein sequence of receptors. Mol Pharmaceutics. 2009;6(3):825‐835.10.1021/mp800102c19281186 doi:10.1021/mp800102c PubMed
198. Speck‐Planche A , Kleandrova VV , Luan F , Cordeiro MN . Chemoinformatics in multi‐target drug discovery for anti‐cancer therapy: in silico design of potent and versatile anti‐brain tumor agents. Anticancer Agents Med Chem. 2012;12(6):678‐685.2204399510.2174/187152012800617722 doi:10.2174/187152012800617722 PubMed
199. Speck‐Planche A , Cordeiro M . Fragment‐based in silico modeling of multi‐target inhibitors against breast cancer‐related proteins. Mol Divers. 2017;21(3):511‐523.2819462710.1007/s11030-017-9731-1 doi:10.1007/s11030-017-9731-1 PubMed
200. Dahl GE , Jaitly N , Salakhutdinov R Multi‐task neural networks for QSAR predictions. arXiv preprint arXiv:14061231. 2014.
201. Schlessinger A , Abagyan R , Carlson HA , Dang KK , Guinney J , Cagan RL . Multi‐targeting drug community challenge. Cell Chemical Biology. 2017;24(12):1434‐1435.2927270010.1016/j.chembiol.2017.12.006 doi:10.1016/j.chembiol.2017.12.006 PubMed
202. Cichonska A , Ravikumar B , Allaway RJ , et al. Crowdsourced mapping extends the target space of kinase inhibitors. bioRxiv. 2020. 10.1101/2019.12.31.891812 doi:10.1101/2019.12.31.891812
203. Xu Y , Ma J , Liaw A , Sheridan RP , Svetnik V . Demystifying multitask deep neural networks for quantitative structure‐activity relationships. J Chem Inf Model. 2017;57(10):2490‐2504.2887286910.1021/acs.jcim.7b00087 doi:10.1021/acs.jcim.7b00087 PubMed
204. Zakharov AV , Zhao T , Nguyen DT , et al. Novel consensus architecture to improve performance of large‐scale multitask deep learning QSAR Models. J Chem Inf Model. 2019;59(11):4613‐4624.3158427010.1021/acs.jcim.9b00526PMC8381874 doi:10.1021/acs.jcim.9b00526 PubMed
205. Kwon S , Bae H , Jo J , Yoon S . Comprehensive ensemble in QSAR prediction for drug discovery. BMC Bioinformatics. 2019;20(1):521.3165554510.1186/s12859-019-3135-4PMC6815455 doi:10.1186/s12859-019-3135-4 PubMed
206. Cheng T , Li Q , Wang Y , Bryant SH . Binary classification of aqueous solubility using support vector machines with reduction and recombination feature selection. J Chem Inf Model. 2011;51(2):229‐236.2121422410.1021/ci100364aPMC3047290 doi:10.1021/ci100364a PubMed
207. Lind P , Maltseva T . Support vector machines for the estimation of aqueous solubility. J Chem Inf Comput Sci. 2003;43(6):1855‐1859.1463243310.1021/ci034107s doi:10.1021/ci034107s PubMed
208. Lusci A , Pollastri G , Baldi P . Deep architectures and deep learning in chemoinformatics: the prediction of aqueous solubility for drug‐like molecules. J Chem Inf Model. 2013;53(7):1563‐1575.2379555110.1021/ci400187yPMC3739985 doi:10.1021/ci400187y PubMed
209. Coley CW , Barzilay R , Green WH , Jaakkola TS , Jensen KF . Convolutional embedding of attributed molecular graphs for physical property prediction. J Chem Inf Model. 2017;57(8):1757‐1772.2869668810.1021/acs.jcim.6b00601 doi:10.1021/acs.jcim.6b00601 PubMed
210. Boobier S , Osbourn A , Mitchell JBO . Can human experts predict solubility better than computers? J Cheminform. 2017;9(1):63.2923889110.1186/s13321-017-0250-yPMC5729181 doi:10.1186/s13321-017-0250-y PubMed
211. Shaik B , Gupta R , Louis B , Agrawal VK . Prediction of permeability of drug‐like compounds across polydimethylsiloxane membranes by machine learning methods. J Pharm Investig. 2015;45(5):461‐473.
212. Chi CT , Lee MH , Weng CF , Leong MK . In silico prediction of PAMPA effective permeability using a two‐QSAR approach. Int J Mol Sci. 2019;20(13):3170.10.3390/ijms20133170PMC665183731261723 doi:10.3390/ijms20133170 PubMed
213. Brocke SA , Degen A , MacKerell AD , Dutagaci B , Feig M . Prediction of membrane permeation of drug molecules by combining an implicit membrane model with machine learning. J Chem Inf Model. 2019;59(3):1147‐1162.3054045910.1021/acs.jcim.8b00648PMC6433486 doi:10.1021/acs.jcim.8b00648 PubMed
214. Cheng T , Zhao Y , Li X , et al. Computation of octanol‐water partition coefficients by guiding an additive model with knowledge. J Chem Inf Model. 2007;47(6):2140‐2148.1798586510.1021/ci700257y doi:10.1021/ci700257y PubMed
215. Zhang H , Xiang ML , Ma CY , et al. Three‐class classification models of logS and logP derived by using GA‐CG‐SVM approach. Mol Divers. 2009;13(2):261‐268.1918463010.1007/s11030-009-9108-1 doi:10.1007/s11030-009-9108-1 PubMed
216. Liao Q , Yao J , Yuan S . SVM approach for predicting LogP. Mol Divers. 2006;10(3):301‐309.1703153410.1007/s11030-006-9036-2 doi:10.1007/s11030-006-9036-2 PubMed
217. Tetko IV , Poda GI . Application of ALOGPS 2.1 to predict log D distribution coefficient for Pfizer proprietary compounds. J Med Chem. 2004;47(23):5601‐5604.1550915610.1021/jm049509l doi:10.1021/jm049509l PubMed
218. Tetko IV , Tanchuk VY , Villa AE . Prediction of n‐octanol/water partition coefficients from PHYSPROP database using artificial neural networks and E‐state indices. J Chem Inf Comput Sci. 2001;41(5):1407‐1421.1160404210.1021/ci010368v doi:10.1021/ci010368v PubMed
219. Bruneau P , McElroy NR . logD7.4 modeling using Bayesian regularized neural networks. Assessment and correction of the errors of prediction. J Chem Inf Model. 2006;46(3):1379‐1387.1671175710.1021/ci0504014 doi:10.1021/ci0504014 PubMed
220. Kola I , Landis J . Can the pharmaceutical industry reduce attrition rates? Nat Rev Drug Discovery. 2004;3(8):711‐716.1528673710.1038/nrd1470 doi:10.1038/nrd1470 PubMed
221. Bhhatarai B , Walters WP , Hop C , Lanza G , Ekins S . Opportunities and challenges using artificial intelligence in ADME/Tox. Nat Mater. 2019;18(5):418‐422.3100080110.1038/s41563-019-0332-5PMC6594826 doi:10.1038/s41563-019-0332-5 PubMed
222. Shen J , Cheng F , Xu Y , Li W , Tang Y . Estimation of ADME properties with substructure pattern recognition. J Chem Inf Model. 2010;50(6):1034‐1041.2057872710.1021/ci100104j doi:10.1021/ci100104j PubMed
223. Yang H , Sun L , Li W , Liu G , Tang Y . In silico prediction of chemical toxicity for drug design using machine learning methods and structural alerts. Front Chem. 2018;6:30.2951599310.3389/fchem.2018.00030PMC5826228 doi:10.3389/fchem.2018.00030 PubMed
224. Huang R , Xia M , Nguyen D‐T , et al. Tox21Challenge to build predictive models of nuclear receptor and stress response pathways as mediated by exposure to environmental chemicals and drugs. Front Environ Sci. 2016;3(85). 10.3389/fenvs.2015.00085 doi:10.3389/fenvs.2015.00085
225. Irwin BWJ , Levell JR , Whitehead TM , Segall MD , Conduit GJ . Practical applications of deep learning to impute heterogeneous drug discovery data. J Chem Inf Model. 2020;60(6):2848‐2857.3247851710.1021/acs.jcim.0c00443 doi:10.1021/acs.jcim.0c00443 PubMed
226. Wang Y , Huang L , Jiang S , et al. Capsule networks showed excellent performance in the classification of hERG blockers/nonblockers. Front Pharmacol. 2019;10:1631.3206384910.3389/fphar.2019.01631PMC6997788 doi:10.3389/fphar.2019.01631 PubMed
227. Tao L , Zhang P , Qin C , et al. Recent progresses in the exploration of machine learning methods as in‐silico ADME prediction tools. Adv Drug Deliv Rev. 2015;86:83‐100.2603706810.1016/j.addr.2015.03.014 doi:10.1016/j.addr.2015.03.014 PubMed
228. Bhhatarai B , Walters WP , Hop CECA , Lanza G , Ekins S . Opportunities and challenges using artificial intelligence in ADME/Tox. Nat Mater. 2019;18(5):418‐422.3100080110.1038/s41563-019-0332-5PMC6594826 doi:10.1038/s41563-019-0332-5 PubMed
229. Zang Q , Mansouri K , Williams AJ , et al. In silico prediction of physicochemical properties of environmental chemicals using molecular fingerprints and machine learning. J Chem Inf Model. 2017;57(1):36‐49.2800689910.1021/acs.jcim.6b00625PMC6131700 doi:10.1021/acs.jcim.6b00625 PubMed
230. Alqahtani S . In silico ADME‐Tox modeling: progress and prospects. Expert Opin Drug Metab Toxicol. 2017;13(11):1147‐1158.2898850610.1080/17425255.2017.1389897 doi:10.1080/17425255.2017.1389897 PubMed
231. Wenzel J , Matter H , Schmidt F . Predictive multitask deep neural network models for ADME‐Tox properties: learning from large data sets. J Chem Inf Model. 2019;59(3):1253‐1268.3061582810.1021/acs.jcim.8b00785 doi:10.1021/acs.jcim.8b00785 PubMed
232. Göller AH , Kuhnke L , Montanari F , et al. Bayer's in silico ADMET platform: a journey of machine learning over the past two decades. Drug Discovery Today. 2020;25:1702‐1709.3265230910.1016/j.drudis.2020.07.001 doi:10.1016/j.drudis.2020.07.001 PubMed
233. Schneider G . Future de novo drug design. Mol Inf. 2014;33(6‐7):397‐402.10.1002/minf.20140003427485977 doi:10.1002/minf.201400034 PubMed
234. Struble TJ , Alvarez JC , Brown SP , et al. Current and future roles of artificial intelligence in medicinal chemistry synthesis. J Med Chem. 2020;63:8667‐8682.3224315810.1021/acs.jmedchem.9b02120PMC7457232 doi:10.1021/acs.jmedchem.9b02120 PubMed
235. Schneider G . Generative models for artificially‐intelligent molecular design. Mol Inf. 2018;37(1‐2):1880131.10.1002/minf.20188013129442446 doi:10.1002/minf.201880131 PubMed
236. Kadurin A , Aliper A , Kazennov A , et al. The cornucopia of meaningful leads: applying deep adversarial autoencoders for new molecule development in oncology. Oncotarget. 2017;8(7):10883‐10890.2802964410.18632/oncotarget.14073PMC5355231 doi:10.18632/oncotarget.14073 PubMed
237. Kadurin A , Nikolenko S , Khrabrov K , Aliper A , Zhavoronkov A . druGAN: an advanced generative adversarial autoencoder model for de novo generation of new molecules with desired molecular properties in silico. Mol Pharmaceutics. 2017;14(9):3098‐3104.10.1021/acs.molpharmaceut.7b0034628703000 doi:10.1021/acs.molpharmaceut.7b00346 PubMed
238. Polykovskiy D , Zhebrak A , Vetrov D , et al. Entangled conditional adversarial autoencoder for de novo drug discovery. Mol Pharmaceutics. 2018;15(10):4398‐4405.10.1021/acs.molpharmaceut.8b0083930180591 doi:10.1021/acs.molpharmaceut.8b00839 PubMed
239. Gómez‐Bombarelli R , Wei JN , Duvenaud D , et al. Automatic chemical design using a data‐driven continuous representation of molecules. ACS Cent Sci. 2018;4(2):268‐276.2953202710.1021/acscentsci.7b00572PMC5833007 doi:10.1021/acscentsci.7b00572 PubMed
240. Lim J , Ryu S , Kim JW , Kim WY . Molecular generative model based on conditional variational autoencoder for de novo molecular design. J Cheminform. 2018;10(1):31.2999527210.1186/s13321-018-0286-7PMC6041224 doi:10.1186/s13321-018-0286-7 PubMed
241. Olivecrona M , Blaschke T , Engkvist O , Chen H . Molecular de‐novo design through deep reinforcement learning. J Cheminform. 2017;9(1):48.2908608310.1186/s13321-017-0235-xPMC5583141 doi:10.1186/s13321-017-0235-x PubMed
242. Segler MHS , Kogej T , Tyrchan C , Waller MP . Generating focused molecule libraries for drug discovery with recurrent neural networks. ACS Cent Sci. 2018;4(1):120‐131.2939218410.1021/acscentsci.7b00512PMC5785775 doi:10.1021/acscentsci.7b00512 PubMed
243. Yuan W , Jiang D , Nambiar DK , et al. Chemical space mimicry for drug discovery. J Chem Inf Model. 2017;57(4):875‐882.2825719110.1021/acs.jcim.6b00754PMC5802964 doi:10.1021/acs.jcim.6b00754 PubMed
244. Merk D , Grisoni F , Friedrich L , Schneider G . Tuning artificial intelligence on the de novo design of natural‐product‐inspired retinoid X receptor modulators. Communications Chemistry. 2018;1(1):68.
245. Popova M , Isayev O , Tropsha A . Deep reinforcement learning for de novo drug design. Sci Adv. 2018;4(7):eaap7885.3005098410.1126/sciadv.aap7885PMC6059760 doi:10.1126/sciadv.aap7885 PubMed
246. Henderson P , Islam R , Bachman P , Pineau J , Precup D , Meger D . Deep reinforcement learning that matters. Paper presented at: Thirty‐Second AAAI Conference on Artificial Intelligence, 2018.
247. Putin E , Asadulaev A , Ivanenkov Y , et al. Reinforced adversarial neural computer for de novo molecular design. J Chem Inf Model. 2018;58(6):1194‐1204.2976202310.1021/acs.jcim.7b00690 doi:10.1021/acs.jcim.7b00690 PubMed
248. Sanchez‐Lengeling B , Outeiral C , Guimaraes GL , Aspuru‐Guzik A . Optimizing distributions over molecular space. An objective‐reinforced generative adversarial network for inverse‐design chemistry (ORGANIC). ChemRxiv. 2017. 10.26434/chemrxiv.5309668.v3 doi:10.26434/chemrxiv.5309668.v3
249. Gupta A , Muller AT , Huisman BJH , Fuchs JA , Schneider P , Schneider G . Generative recurrent networks for de novo drug design. Mol Inf. 2018;37(1–2):1700111.10.1002/minf.201700111PMC583694329095571 doi:10.1002/minf.201700111 PubMed
250. Awale M , Sirockin F , Stiefl N , Reymond JL . Drug analogs from fragment‐based long short‐term memory generative neural networks. J Chem Inf Model. 2019;59(4):1347‐1356.3090891310.1021/acs.jcim.8b00902 doi:10.1021/acs.jcim.8b00902 PubMed
251. Liu B , Ramsundar B , Kawthekar P , et al. Retrosynthetic reaction prediction using neural sequence‐to‐sequence models. ACS Cent Sci. 2017;3(10):1103‐1113.2910492710.1021/acscentsci.7b00303PMC5658761 doi:10.1021/acscentsci.7b00303 PubMed
252. Noel OB , Andrew D . DeepSMILES: an adaptation of SMILES for use in machine‐learning of chemical structures. ChemRxiv. 2018. 10.26434/chemrxiv.7097960.v1 doi:10.26434/chemrxiv.7097960.v1
253. Yoshikawa N , Terayama K , Sumita M , Homma T , Oono K , Tsuda KJCL . Population‐based de novo molecule generation, using grammatical evolution. Chem Lett. 2018;47(11):1431‐1434.
254. Pope PE , Kolouri S , Rostami M , Martin CE , Hoffmann H . Discovering molecular functional groups using graph convolutional neural networks. ArXiv. 2018. abs/1812.00265.
255. You J , Liu B , Ying R , Pande V , Leskovec J . Graph convolutional policy network for goal‐directed molecular graph generation. In: Proceedings of the 32nd International Conference on Neural Information Processing Systems. Montréal, Canada; 2018:6412‐6422.
256. De Cao N , Kipf T . MolGAN: an implicit generative model for small molecular graphs. ArXiv Preprint. 2018;arXiv:180511973.
257. Maziarka Ł , Pocha A , Kaczmarczyk J , Rataj K , Danel T , Warchoł M . Mol‐CycleGAN: a generative model for molecular optimization. J Cheminform. 2020;12(1):1‐18. 10.1186/s13321-019-0404-1 33431006PMC6950853 doi:10.1186/s13321-019-0404-1 PubMed
258. Liu Q , Allamanis M , Brockschmidt M , Gaunt A . Constrained graph variational autoencoders for molecule design. In: Proceedings of the 32nd International Conference on Neural Information Processing Systems. Montréal, Canada; 2018:7806‐7815.
259. Skalic M , Jiménez J , Sabbadin D , de Fabritiis G . Shape‐based generative modeling for de novo drug design. J Chem Inf Model. 2019;59(3):1205‐1214.3076236410.1021/acs.jcim.8b00706 doi:10.1021/acs.jcim.8b00706 PubMed
260. Button A , Merk D , Hiss JA , Schneider G . Automated de novo molecular design by hybrid machine intelligence and rule‐driven chemical synthesis. Nat Mach Intell. 2019;1(7):307‐315.
261. Méndez‐Lucio O , Baillif B , Clevert D‐A , Rouquié D , Wichard J . De novo generation of hit‐like molecules from gene expression signatures using artificial intelligence. Nat Commun. 2020;11(1):10.3190040810.1038/s41467-019-13807-wPMC6941972 doi:10.1038/s41467-019-13807-w PubMed
262. Deep learning enables rapid identification of potent DDR1 kinase inhibitors. Overview of attention for article published in Nature Biotechnology, September 2019. Altmetric. https://www.altmetric.com/details/65797801. Accessed October 2, 2020.10.1038/s41587-019-0224-x31477924 doi:10.1038/s41587-019-0224-x PubMed
263. Zhavoronkov A , Ivanenkov YA , Aliper A , et al. Deep learning enables rapid identification of potent DDR1 kinase inhibitors. Nature Biotechnol. 2019;37(9):1038‐1040.3147792410.1038/s41587-019-0224-x doi:10.1038/s41587-019-0224-x PubMed
264. Hanson SM , Georghiou G , Thakur MK , et al. What makes a kinase promiscuous for inhibitors? Cell Chem Biol. 2019;26(3):390‐399.3061295110.1016/j.chembiol.2018.11.005PMC6632086 doi:10.1016/j.chembiol.2018.11.005 PubMed
265. Walters WP , Murcko M . Assessing the impact of generative AI on medicinal chemistry. Nature Biotechnol. 2020;38(2):143‐145.3200183410.1038/s41587-020-0418-2 doi:10.1038/s41587-020-0418-2 PubMed
266. Rampasek L , Hidru D , Smirnov P , Haibe‐Kains B , Goldenberg A . Dr.VAE: improving drug response prediction via modeling of drug perturbation effects. Bioinformatics. 2019;35(19):3743‐3751.3085084610.1093/bioinformatics/btz158PMC6761940 doi:10.1093/bioinformatics/btz158 PubMed
267. Geeleher P , Cox NJ , Huang RS . Clinical drug response can be predicted using baseline gene expression levels and in vitro drug sensitivity in cell lines. Genome Biol. 2014;15(3):R47.2458083710.1186/gb-2014-15-3-r47PMC4054092 doi:10.1186/gb-2014-15-3-r47 PubMed
268. Ding Z , Zu S , Gu J . Evaluating the molecule‐based prediction of clinical drug responses in cancer. Bioinformatics. 2016;32(19):2891‐2895.2735469410.1093/bioinformatics/btw344 doi:10.1093/bioinformatics/btw344 PubMed
269. Iorio F , Knijnenburg TA , Vis DJ , et al. A landscape of pharmacogenomic interactions in cancer. Cell. 2016;166(3):740‐754.2739750510.1016/j.cell.2016.06.017PMC4967469 doi:10.1016/j.cell.2016.06.017 PubMed
270. Tan M , Ozgul OF , Bardak B , Eksioglu I , Sabuncuoglu S . Drug response prediction by ensemble learning and drug‐induced gene expression signatures. Genomics. 2019;111(5):1078‐1088.3153390010.1016/j.ygeno.2018.07.002 doi:10.1016/j.ygeno.2018.07.002 PubMed
271. Graim K , Friedl V , Houlahan KE , Stuart JM . PLATYPUS: a multiple‐view learning predictive framework for cancer drug sensitivity prediction. Pac Symp Biocomput. 2019;24:136‐147.30864317PMC6417802 PubMed
272. Sharifi‐Noghabi H , Zolotareva O , Collins CC , Ester M . MOLI: multi‐omics late integration with deep neural networks for drug response prediction. Bioinformatics. 2019;35(14):i501‐i509.3151070010.1093/bioinformatics/btz318PMC6612815 doi:10.1093/bioinformatics/btz318 PubMed
273. Dincer AB , Celik S , Hiranuma N , Lee S‐I . DeepProfile: deep learning of cancer molecular profiles for precision medicine. bioRxiv. 2018:278739. 10.1101/278739 doi:10.1101/278739
274. Li M , Wang Y , Zheng R , et al. DeepDSC: a deep learning method to predict drug sensitivity of cancer cell lines. IEEE/ACM Trans Comput Biol Bioinform. 2019:1. 10.1109/tcbb.2019.2919581 31150344 doi:10.1109/tcbb.2019.2919581 PubMed
275. Ding MQ , Chen L , Cooper GF , Young JD , Lu X . Precision oncology beyond targeted therapy: combining omics data with machine learning matches the majority of cancer cells to effective therapeutics. Mol Cancer Res. 2018;16(2):269‐278.2913358910.1158/1541-7786.MCR-17-0378PMC5821274 doi:10.1158/1541-7786.MCR-17-0378 PubMed
276. Iwata M , Yuan L , Zhao Q , et al. Predicting drug‐induced transcriptome responses of a wide range of human cell lines by a novel tensor‐train decomposition algorithm. Bioinformatics. 2019;35(14):i191‐i199.3151066310.1093/bioinformatics/btz313PMC6612872 doi:10.1093/bioinformatics/btz313 PubMed
277. Tan M . Prediction of anti‐cancer drug response by kernelized multi‐task learning. Artif Intell Med. 2016;73:70‐77.2792638210.1016/j.artmed.2016.09.004 doi:10.1016/j.artmed.2016.09.004 PubMed
278. Ammad‐Ud‐Din M , Khan SA , Wennerberg K , Aittokallio T . Systematic identification of feature combinations for predicting drug response with Bayesian multi‐view multi‐task linear regression. Bioinformatics. 2017;33(14):i359‐i368.2888199810.1093/bioinformatics/btx266PMC5870540 doi:10.1093/bioinformatics/btx266 PubMed
279. Greco WR , Faessel H , Levasseur L . The search for cytotoxic synergy between anticancer agents: a case of Dorothy and the ruby slippers? J Natl Cancer Inst. 1996;88(11):699‐700.863701810.1093/jnci/88.11.699 doi:10.1093/jnci/88.11.699 PubMed
280. Roell KR , Reif DM , Motsinger‐Reif AA . An Introduction to terminology and methodology of chemical synergy‐perspectives from across disciplines. Front Pharmacol. 2017;8:158.2847376910.3389/fphar.2017.00158PMC5397413 doi:10.3389/fphar.2017.00158 PubMed
281. Gibbs BK , Sourbier C . Detecting the potential pharmacological synergy of drug combination by viability assays in vitro. Methods Mol Biol (Clifton, NJ). 2018;1709:129‐137.10.1007/978-1-4939-7477-1_1029177656 doi:10.1007/978-1-4939-7477-1_10 PubMed
282. Mayr A , Klambauer G , Unterthiner T , Hochreiter S . DeepTox: toxicity prediction using deep learning. Front Environ Sci. 2016;3(80). 10.3389/fenvs.2015.00080 doi:10.3389/fenvs.2015.00080
283. Madani Tonekaboni SA , Soltan Ghoraie L , Manem VSK , Haibe‐Kains B . Predictive approaches for drug combination discovery in cancer. Brief Bioinform. 2018;19(2):263‐276.2788143110.1093/bib/bbw104PMC6018991 doi:10.1093/bib/bbw104 PubMed
284. Preuer K , Lewis RPI , Hochreiter S , Bender A , Bulusu KC , Klambauer G . DeepSynergy: predicting anti‐cancer drug synergy with deep learning. Bioinformatics. 2018;34(9):1538‐1546.2925307710.1093/bioinformatics/btx806PMC5925774 doi:10.1093/bioinformatics/btx806 PubMed
285. Malyutina A , Majumder MM , Wang W , Pessia A , Heckman CA , Tang J . Drug combination sensitivity scoring facilitates the discovery of synergistic and efficacious drug combinations in cancer. PLOS Comput Biol. 2019;15(5):e1006752.3110786010.1371/journal.pcbi.1006752PMC6544320 doi:10.1371/journal.pcbi.1006752 PubMed
286. Sidorov P , Naulaerts S , Ariey‐Bonnet J , Pasquier E , Ballester PJ . Predicting synergism of cancer drug combinations using NCI‐ALMANAC data. Front Chem. 2019;7:509.3138035210.3389/fchem.2019.00509PMC6646421 doi:10.3389/fchem.2019.00509 PubMed
287. Celebi R , Bear Don't Walk O 4th , Movva R , Alpsoy S , Dumontier M . In‐silico prediction of synergistic anti‐cancer drug combinations using multi‐omics data. Sci Rep. 2019;9(1):8949.3122210910.1038/s41598-019-45236-6PMC6586895 doi:10.1038/s41598-019-45236-6 PubMed
288. Singh H , Rana PS , Singh U . Prediction of drug synergy in cancer using ensemble‐based machine learning techniques. Mod Phys Lett B. 2018;32(11):1850132.
289. Gallagher PJ , Castro V , Fava M , et al. Antidepressant response in patients with major depression exposed to NSAIDs: a pharmacovigilance study. Am J Psychiatry. 2012;169(10):1065‐1072.2303238610.1176/appi.ajp.2012.11091325PMC3787520 doi:10.1176/appi.ajp.2012.11091325 PubMed
290. Mason DJ , Eastman RT , Lewis RPI , Stott IP , Guha R , Bender A . Using machine learning to predict synergistic antimalarial compound combinations with novel structures. Front Pharmacol. 2018;9:1096.3033374810.3389/fphar.2018.01096PMC6176478 doi:10.3389/fphar.2018.01096 PubMed
291. Mason DJ , Stott I , Ashenden S , et al. Prediction of antibiotic interactions using descriptors derived from molecular structure. J Med Chem. 2017;60(9):3902‐3912.2838390210.1021/acs.jmedchem.7b00204 doi:10.1021/acs.jmedchem.7b00204 PubMed
292. Ianevski A , Giri AK , Gautam P , et al. Prediction of drug combination effects with a minimal set of experiments. Nat Mach Intell. 2019;1(12):568‐577.3236872110.1038/s42256-019-0122-4PMC7198051 doi:10.1038/s42256-019-0122-4 PubMed
293. Ding P , Yin R , Luo J , Kwoh CK . Ensemble prediction of synergistic drug combinations incorporating biological, chemical, pharmacological, and network knowledge. IEEE J Biomed Health Inform. 2019;23(3):1336‐1345.2999440810.1109/JBHI.2018.2852274 doi:10.1109/JBHI.2018.2852274 PubMed
294. Zhang C , Yan G . Synergistic drug combinations prediction by integrating pharmacological data. Synth Syst Biotechnol. 2019;4(1):67‐72.3082047810.1016/j.synbio.2018.10.002PMC6370570 doi:10.1016/j.synbio.2018.10.002 PubMed
295. Vilar S , Harpaz R , Uriarte E , Santana L , Rabadan R , Friedman C . Drug‐drug interaction through molecular structure similarity analysis. J Am Med Inform Assoc. 2012;19(6):1066‐1074.2264769010.1136/amiajnl-2012-000935PMC3534468 doi:10.1136/amiajnl-2012-000935 PubMed
296. Zitnik M , Agrawal M , Leskovec J . Modeling polypharmacy side effects with graph convolutional networks. Bioinformatics. 2018;34(13):i457‐i466.2994999610.1093/bioinformatics/bty294PMC6022705 doi:10.1093/bioinformatics/bty294 PubMed
297. Ryu JY , Kim HU , Lee SY . Deep learning improves prediction of drug–drug and drug–food interactions. Proc Natl Acad Sci. 2018;115(18):E4304‐E4311.2966622810.1073/pnas.1803294115PMC5939113 doi:10.1073/pnas.1803294115 PubMed
298. Zheng Y , Peng H , Zhang X , Zhao Z , Yin J , Li J . Predicting adverse drug reactions of combined medication from heterogeneous pharmacologic databases. BMC Bioinformatics. 2018;19(19):517.3059806510.1186/s12859-018-2520-8PMC6311930 doi:10.1186/s12859-018-2520-8 PubMed
299. Lee CY , Chen YP . Prediction of drug adverse events using deep learning in pharmaceutical discovery. Brief Bioinform. 2020. 10.1093/bib/bbaa040 32349125 doi:10.1093/bib/bbaa040 PubMed
300. Shankar S , Bhandari I , Okou DT , Srinivasa G , Athri P . Predicting adverse drug reactions of two‐drug combinations using structural and transcriptomic drug representations to train an artificial neural network. Chem Biol Drug Des. 2020:cbdd.13802 10.1111/cbdd.1380233006799 doi:10.1111/cbdd.13802 PubMed
301. Avorn J . The $2.6 billion pill—methodologic and policy considerations. N Engl J Med. 2015;372(20):1877‐1879.2597004910.1056/NEJMp1500848 doi:10.1056/NEJMp1500848 PubMed
302. Mullard A . FDA drug approvals. Nat Rev Drug Discovery. 2016;2017 16(2):73‐76.10.1038/nrd.2017.1428148938 doi:10.1038/nrd.2017.14 PubMed
303. Tan SY , Grimes S . Paul Ehrlich (1854–1915): man with the magic bullet. Singapore Med J. 2010;51(11):842‐843.21140107 PubMed
304. Greene JA , Loscalzo J . Putting the patient back together—social medicine, network medicine, and the limits of reductionism. N Engl J Med. 2017;377(25):2493‐2499.2926227710.1056/NEJMms1706744 doi:10.1056/NEJMms1706744 PubMed
305. Gottlieb A , Stein GY , Ruppin E , Sharan R . PREDICT: a method for inferring novel drug indications with application to personalized medicine. Mol Syst Biol. 2011;7:496.2165467310.1038/msb.2011.26PMC3159979 doi:10.1038/msb.2011.26 PubMed
306. Liu Z , Guo F , Gu J , et al. Similarity‐based prediction for anatomical therapeutic chemical classification of drugs by integrating multiple data sources. Bioinformatics. 2015;31(11):1788‐1795.2563881010.1093/bioinformatics/btv055 doi:10.1093/bioinformatics/btv055 PubMed
307. Napolitano F , Zhao Y , Moreira VM , et al. Drug repositioning: a machine‐learning approach through data integration. J Cheminform. 2013;5(1):30.2380001010.1186/1758-2946-5-30PMC3704944 doi:10.1186/1758-2946-5-30 PubMed
308. Masoudi‐Sobhanzadeh Y , Omidi Y , Amanlou M , Masoudi‐Nejad A . DrugR+: a comprehensive relational database for drug repurposing, combination therapy, and replacement therapy. Comput Biol Med. 2019;109:254‐262.3109608910.1016/j.compbiomed.2019.05.006 doi:10.1016/j.compbiomed.2019.05.006 PubMed
309. Ravikumar B , Timonen S , Alam Z , Parri E , Wennerberg K , Aittokallio T . Chemogenomic analysis of the druggable kinome and its application to repositioning and lead identification studies. Cell Chem Biol. 2019;26(11):1608‐1622.3152162210.1016/j.chembiol.2019.08.007 doi:10.1016/j.chembiol.2019.08.007 PubMed
310. Oh M , Ahn J , Yoon Y . A network‐based classification model for deriving novel drug‐disease associations and assessing their molecular actions. PLOS One. 2014;9(10):e111668.2535691010.1371/journal.pone.0111668PMC4214731 doi:10.1371/journal.pone.0111668 PubMed
311. Zheng Y , Peng H , Zhang X , Zhao Z , Gao X , Li J . Old drug repositioning and new drug discovery through similarity learning from drug‐target joint feature spaces. BMC Bioinformatics. 2019;20(suppl 23):605.3188182910.1186/s12859-019-3238-yPMC6933655 doi:10.1186/s12859-019-3238-y PubMed
312. Zhang P , Agarwal P , Obradovic Z . Computational Drug Repositioning by Ranking and Integrating Multiple Data Sources. Berlin, Heidelberg, Germany: Springer Verlag; 2013.
313. Xuan P , Zhao L , Zhang T , Ye Y , Zhang Y . Inferring drug‐related diseases based on convolutional neural network and gated recurrent unit. Molecules (Basel, Switzerland). 2019;24(15):2712.10.3390/molecules24152712PMC669644331349692 doi:10.3390/molecules24152712 PubMed
314. Aliper A , Plis S , Artemov A , Ulloa A , Mamoshina P , Zhavoronkov A . Deep learning applications for predicting pharmacological properties of drugs and drug repurposing using transcriptomic data. Mol Pharmaceutics. 2016;13(7):2524‐2530.10.1021/acs.molpharmaceut.6b00248PMC496526427200455 doi:10.1021/acs.molpharmaceut.6b00248 PubMed
315. Wu G , Liu J , Yue X . Prediction of drug‐disease associations based on ensemble meta paths and singular value decomposition. BMC Bioinformatics. 2019;20(suppl 3):134.3092585810.1186/s12859-019-2644-5PMC6439991 doi:10.1186/s12859-019-2644-5 PubMed
316. Zhao D , Wang J , Sang S , Lin H , Wen J , Yang C . Relation path feature embedding based convolutional neural network method for drug discovery. BMC Med Inform Decis Mak. 2019;19(suppl 2):59.3096159910.1186/s12911-019-0764-5PMC6454669 doi:10.1186/s12911-019-0764-5 PubMed
317. Jiang HJ , Huang YA , You ZH . Predicting drug‐disease associations via using Gaussian interaction profile and Kernel‐based autoencoder. BioMed Res Int. 2019;2019:1‐11. 10.1155/2019/2426958 PMC673262231534955 doi:10.1155/2019/2426958 PubMed
318. Lee M , Kim H , Joe H , Kim H‐G . Multi‐channel PINN: investigating scalable and transferable neural networks for drug discovery. J Cheminform. 2019;11(1):46.3128996310.1186/s13321-019-0368-1PMC6617572 doi:10.1186/s13321-019-0368-1 PubMed
319. Wang R , Li S , Cheng L , Wong MH , Leung KS . Predicting associations among drugs, targets and diseases by tensor decomposition for drug repositioning. BMC Bioinformatics. 2019;20(suppl 26):628.3183900810.1186/s12859-019-3283-6PMC6912989 doi:10.1186/s12859-019-3283-6 PubMed
320. Bahi M , Batouche M . Drug‐Target Interaction Prediction in Drug Repositioning Based on Deep Semi‐Supervised Learning. Cham, Switzerland: Springer; 2018.
321. deAndres‐Galiana EJ , Bea G , Fernandez‐Martinez JL , Saligan LN . Analysis of defective pathways and drug repositioning in Multiple Sclerosis via machine learning approaches. Comput Biol Med. 2019;115:103492.3162701710.1016/j.compbiomed.2019.103492 doi:10.1016/j.compbiomed.2019.103492 PubMed
322. Xia Z , Wu LY , Zhou X , Wong ST . Semi‐supervised drug‐protein interaction prediction from heterogeneous biological spaces. BMC Syst Biol. 2010;4(suppl 2):S6.10.1186/1752-0509-4-S2-S6PMC298269320840733 doi:10.1186/1752-0509-4-S2-S6 PubMed
323. Chen H , Zhang Z . A semi‐supervised method for drug‐target interaction prediction with consistency in networks. PLOS One. 2013;8(5):e62975.2366755310.1371/journal.pone.0062975PMC3646965 doi:10.1371/journal.pone.0062975 PubMed
324. Nam Y , Kim M , Chang H‐S , Shin H . Drug repurposing with network reinforcement. BMC Bioinformatics. 2019;20(13):383.3133733310.1186/s12859-019-2858-6PMC6651901 doi:10.1186/s12859-019-2858-6 PubMed
325. Yan XY , Zhang SW , Zhang SY . Prediction of drug‐target interaction by label propagation with mutual interaction information derived from heterogeneous network. Mol BioSyst. 2016;12(2):520‐531.2667553410.1039/c5mb00615e doi:10.1039/c5mb00615e PubMed
326. Zeng X , Zhu S , Liu X , Zhou Y , Nussinov R , Cheng F . deepDR: a network‐based deep learning approach to in silico drug repositioning. Bioinformatics. 2019;35(24):5191‐5198.3111639010.1093/bioinformatics/btz418PMC6954645 doi:10.1093/bioinformatics/btz418 PubMed
327. Feustel SM , Meissner M , Liesenfeld O . Toxoplasma gondii and the blood‐brain barrier. Virulence. 2012;3(2):182‐192.2246064510.4161/viru.19004PMC3396697 doi:10.4161/viru.19004 PubMed
328. Ballabh P , Braun A , Nedergaard M . The blood‐brain barrier: an overview: structure, regulation, and clinical implications. Neurobiol Dis. 2004;16(1):1‐13.1520725610.1016/j.nbd.2003.12.016 doi:10.1016/j.nbd.2003.12.016 PubMed
329. Doniger S , Hofmann T , Yeh J . Predicting CNS permeability of drug molecules: comparison of neural network and support vector machine algorithms. J Comput Biol. 2002;9(6):849‐864.1261455110.1089/10665270260518317 doi:10.1089/10665270260518317 PubMed
330. Martins IF , Teixeira AL , Pinheiro L , Falcao AO . A Bayesian approach to in silico blood‐brain barrier penetration modeling. J Chem Inf Model. 2012;52(6):1686‐1697.2261259310.1021/ci300124c doi:10.1021/ci300124c PubMed
331. Majumdar S , Basak SC , Lungu CN , Diudea MV , Grunwald GD . Finding needles in a haystack: determining key molecular descriptors associated with the blood‐brain barrier entry of chemical compounds using machine learning. Mol Inf. 2019;38:1800164.10.1002/minf.20180016431322827 doi:10.1002/minf.201800164 PubMed
332. Golmohammadi H , Dashtbozorgi Z , Acree WE Jr . Quantitative structure‐activity relationship prediction of blood‐to‐brain partitioning behavior using support vector machine. Eur J Pharm Sci. 2012;47(2):421‐429.2277154810.1016/j.ejps.2012.06.021 doi:10.1016/j.ejps.2012.06.021 PubMed
333. Yuan Y , Zheng F , Zhan CG . Improved prediction of blood‐brain barrier permeability through machine learning with combined use of molecular property‐based descriptors and fingerprints. AAPS J. 2018;20(3):54.2956457610.1208/s12248-018-0215-8PMC7737623 doi:10.1208/s12248-018-0215-8 PubMed
334. Mente SR , Lombardo F . A recursive‐partitioning model for blood‐brain barrier permeation. J Comput Aided Mol Des. 2005;19(7):465‐481.1633140610.1007/s10822-005-9001-7 doi:10.1007/s10822-005-9001-7 PubMed
335. Zhao YH , Abraham MH , Ibrahim A , et al. Predicting penetration across the blood‐brain barrier from simple descriptors and fragmentation schemes. J Chem Inf Model. 2007;47(1):170‐175.1723826210.1021/ci600312d doi:10.1021/ci600312d PubMed
336. Obrezanova O , Csanyi G , Gola JM , Segall MD . Gaussian processes: a method for automatic QSAR modeling of ADME properties. J Chem Inf Model. 2007;47(5):1847‐1857.1760254910.1021/ci7000633 doi:10.1021/ci7000633 PubMed
337. Andres C , Hutter MC . CNS permeability of drugs predicted by a decision tree. QSAR Comb Sci. 2006;25(4):305‐309.
338. Zhang L , Zhu H , Oprea TI , Golbraikh A , Tropsha A . QSAR modeling of the blood‐brain barrier permeability for diverse organic compounds. Pharm Res. 2008;25(8):1902‐1914.1855321710.1007/s11095-008-9609-0 doi:10.1007/s11095-008-9609-0 PubMed
339. Vilar S , Chakrabarti M , Costanzi S . Prediction of passive blood‐brain partitioning: straightforward and effective classification models based on in silico derived physicochemical descriptors. J Mol Graph Model. 2010;28(8):899‐903.2042721710.1016/j.jmgm.2010.03.010PMC2873098 doi:10.1016/j.jmgm.2010.03.010 PubMed
340. Wang Z , Yang H , Wu Z , et al. In silico prediction of blood‐brain barrier permeability of compounds by machine learning and resampling methods. ChemMedChem. 2018;13(20):2189‐2201.3011051110.1002/cmdc.201800533 doi:10.1002/cmdc.201800533 PubMed
341. Garg P , Verma J . In silico prediction of blood brain barrier permeability: an Artificial Neural Network model. J Chem Inf Model. 2006;46(1):289‐297.1642606410.1021/ci050303i doi:10.1021/ci050303i PubMed
342. Garg P , Dhakne R , Belekar V . Role of breast cancer resistance protein (BCRP) as active efflux transporter on blood‐brain barrier (BBB) permeability. Mol Divers. 2015;19(1):163‐172.2550223410.1007/s11030-014-9562-2 doi:10.1007/s11030-014-9562-2 PubMed
343. Lingineni K , Belekar V , Tangadpalliwar SR , Garg P . The role of multidrug resistance protein (MRP‐1) as an active efflux transporter on blood‐brain barrier (BBB) permeability. Mol Divers. 2017;21(2):355‐365.2805068710.1007/s11030-016-9715-6 doi:10.1007/s11030-016-9715-6 PubMed
344. Geier EG , Schlessinger A , Fan H , et al. Structure‐based ligand discovery for the large‐neutral amino acid transporter 1, LAT‐1. Proc Natl Acad Sci USA. 2013;110(14):5480‐5485.2350925910.1073/pnas.1218165110PMC3619328 doi:10.1073/pnas.1218165110 PubMed
345. Dolghih E , Jacobson MP . Predicting efflux ratios and blood‐brain barrier penetration from chemical structure: combining passive permeability with active efflux by P‐glycoprotein. ACS Chem Neurosci. 2013;4(2):361‐367.2342168710.1021/cn3001922PMC3582291 doi:10.1021/cn3001922 PubMed
346. Hindle SJ , Munji RN , Dolghih E , et al. Evolutionarily conserved roles for blood‐brain barrier xenobiotic transporters in endogenous steroid partitioning and behavior. Cell Rep. 2017;21(5):1304‐1316.2909176810.1016/j.celrep.2017.10.026PMC5774027 doi:10.1016/j.celrep.2017.10.026 PubMed
347. Miller DS . Regulation of ABC transporters blood‐brain barrier: the good, the bad, and the ugly. Adv Cancer Res. 2015;125:43‐70.2564026610.1016/bs.acr.2014.10.002 doi:10.1016/bs.acr.2014.10.002 PubMed
348. Qosa H , Miller DS , Pasinelli P , Trotti D . Regulation of ABC efflux transporters at blood‐brain barrier in health and neurological disorders. Brain Res. 2015;1628(pt B):298‐316.2618775310.1016/j.brainres.2015.07.005PMC4681613 doi:10.1016/j.brainres.2015.07.005 PubMed
349. Gao Z , Chen Y , Cai X , Xu R . Predict drug permeability to blood‐brain‐barrier from clinical phenotypes: drug side effects and drug indications. Bioinformatics. 2017;33(6):901‐908.2799378510.1093/bioinformatics/btw713PMC5860495 doi:10.1093/bioinformatics/btw713 PubMed
350. Miao R , Xia LY , Chen HH , Huang HH , Liang Y . Improved classification of blood‐brain‐barrier drugs using deep learning. Sci Rep. 2019;9(1):8802.3121742410.1038/s41598-019-44773-4PMC6584536 doi:10.1038/s41598-019-44773-4 PubMed
351. Sullivan PF . Puzzling over schizophrenia: schizophrenia as a pathway disease. Nat Med. 2012;18(2):210‐211.2231068710.1038/nm.2670PMC4188922 doi:10.1038/nm.2670 PubMed
352. Stachowiak MK , Kucinski A , Curl R , et al. Schizophrenia: a neurodevelopmental disorder—Integrative genomic hypothesis and therapeutic implications from a transgenic mouse model. Schizophrenia Research. 2013;143(2):367‐376.2323187710.1016/j.schres.2012.11.004 doi:10.1016/j.schres.2012.11.004 PubMed
353. McGrath J , Saha S , Chant D , Welham J . Schizophrenia: a concise overview of incidence, prevalence, and mortality. Epidemiol Rev. 2008;30:67‐76.1848009810.1093/epirev/mxn001 doi:10.1093/epirev/mxn001 PubMed
354. Mathers C . The global burden of disease: 2004 update. World Health Organization; 2008.
355. Kambeitz J , Kambeitz‐Ilankovic L , Leucht S , et al. Detecting neuroimaging biomarkers for schizophrenia: a meta‐analysis of multivariate pattern recognition studies. Neuropsychopharmacology. 2015;40(7):1742‐1751.2560122810.1038/npp.2015.22PMC4915258 doi:10.1038/npp.2015.22 PubMed
356. Hyman SE . Revolution stalled. Sci Transl Med. 2012;4(155):155cm111.10.1126/scitranslmed.300314223052291 doi:10.1126/scitranslmed.3003142 PubMed
357. Palaniyappan L , Deshpande G , Lanka P , et al. Effective connectivity within a triple network brain system discriminates schizophrenia spectrum disorders from psychotic bipolar disorder at the single‐subject level. Schizophr Res. 2019;214:24‐33.2939820710.1016/j.schres.2018.01.006 doi:10.1016/j.schres.2018.01.006 PubMed
358. Schnack HG . Improving individual predictions: machine learning approaches for detecting and attacking heterogeneity in schizophrenia (and other psychiatric diseases). Schizophr Res. 2019;214:34‐42.2907433210.1016/j.schres.2017.10.023 doi:10.1016/j.schres.2017.10.023 PubMed
359. Honnorat N , Dong A , Meisenzahl‐Lechner E , Koutsouleris N , Davatzikos C . Neuroanatomical heterogeneity of schizophrenia revealed by semi‐supervised machine learning methods. Schizophr Res. 2019;214:43‐50.2927473510.1016/j.schres.2017.12.008PMC6013334 doi:10.1016/j.schres.2017.12.008 PubMed
360. Tandon N , Tandon R . Using machine learning to explain the heterogeneity of schizophrenia. Realizing the promise and avoiding the hype. Schizophr Res. 2019;214:70‐75.3150099810.1016/j.schres.2019.08.032 doi:10.1016/j.schres.2019.08.032 PubMed
361. Talpalaru A , Bhagwat N , Devenyi GA , Lepage M , Chakravarty MM . Identifying schizophrenia subgroups using clustering and supervised learning. Schizophr Res. 2019;214:51‐59.3145551810.1016/j.schres.2019.05.044 doi:10.1016/j.schres.2019.05.044 PubMed
362. Mothi SS , Sudarshan M , Tandon N , et al. Machine learning improved classification of psychoses using clinical and biological stratification: update from the bipolar‐schizophrenia network for intermediate phenotypes (B‐SNIP). Schizophr Res. 2019;214:60‐69.2980780410.1016/j.schres.2018.04.037 doi:10.1016/j.schres.2018.04.037 PubMed
363. Hsu KC , Wang FS . Model‐based optimization approaches for precision medicine: a case study in presynaptic dopamine overactivity. PLOS One. 2017;12(6):e0179575.2861441010.1371/journal.pone.0179575PMC5470743 doi:10.1371/journal.pone.0179575 PubMed
364. Yang QX , Wang YX , Li FC , et al. Identification of the gene signature reflecting schizophrenia's etiology by constructing artificial intelligence‐based method of enhanced reproducibility. CNS Neurosci Ther. 2019;25(9):1054‐1063.3135082410.1111/cns.13196PMC6698965 doi:10.1111/cns.13196 PubMed
365. Marunnan SM , Pulikkal BP , Jabamalairaj A , et al. Development of MLR and SVM aided QSAR models to identify common SAR of GABA uptake herbal inhibitors used in the treatment of schizophrenia. Curr Neuropharmacol. 2017;15(8):1085‐1092.2791921110.2174/1567201814666161205131745PMC5725540 doi:10.2174/1567201814666161205131745 PubMed
366. Bain EE , Shafner L , Walling DP , et al. Use of a novel artificial intelligence platform on mobile devices to assess dosing compliance in a phase 2 clinical trial in subjects with schizophrenia. JMIR Mhealth Uhealth. 2017;5(2):e18.2822326510.2196/mhealth.7030PMC5340925 doi:10.2196/mhealth.7030 PubMed
367. Zhao K , So HC . Drug Repositioning for schizophrenia and depression/anxiety disorders: a machine learning approach leveraging expression. Data. IEEE J Biomed Health Inform. 2019;23(3):1304‐1315.3001060310.1109/JBHI.2018.2856535 doi:10.1109/JBHI.2018.2856535 PubMed
368. Yin L , Chau CKL , Sham PC , So HC . Integrating clinical data and imputed transcriptome from GWAS to uncover complex disease subtypes: applications in psychiatry and cardiology. Am J Hum Genet. 2019;105(6):1193‐1212.3178578610.1016/j.ajhg.2019.10.012PMC6904812 doi:10.1016/j.ajhg.2019.10.012 PubMed
369. Chakravarty MM . Guest editorial: Special issue on machine learning in schizophrenia. Schizophr Res. 2019;214:1‐2.3171173210.1016/j.schres.2019.10.044 doi:10.1016/j.schres.2019.10.044 PubMed
370. Fitzpatrick SE , Srivorakiat L , Wink LK , Pedapati EV , Erickson CA . Aggression in autism spectrum disorder: presentation and treatment options. Neuropsychiatr Dis Treat. 2016;12:1525‐1538.2738229510.2147/NDT.S84585PMC4922773 doi:10.2147/NDT.S84585 PubMed
371. Anzulewicz A , Sobota K , Delafield‐Butt JT . Toward the autism motor signature: gesture patterns during smart tablet gameplay identify children with autism. Sci Rep. 2016;6:31107.2755397110.1038/srep31107PMC4995518 doi:10.1038/srep31107 PubMed
372. Obara T , Ishikuro M , Tamiya G , et al. Potential identification of vitamin B6 responsiveness in autism spectrum disorder utilizing phenotype variables and machine learning methods. Sci Rep. 2018;8(1):14840.3028786410.1038/s41598-018-33110-wPMC6172273 doi:10.1038/s41598-018-33110-w PubMed
373. Ekins S , Gerlach J , Zorn KM , Antonio BM , Lin Z , Gerlach A . Repurposing approved drugs as inhibitors of Kv7.1 and Nav1.8 to treat Pitt Hopkins Syndrome. Pharm Res. 2019;36(9):137.3133253310.1007/s11095-019-2671-yPMC6814258 doi:10.1007/s11095-019-2671-y PubMed
374. Heyne HO , Baez‐Nieto D , Iqbal S , et al. Predicting functional effects of missense variants in voltage‐gated sodium and calcium channels. Sci Transl Med. 2020;12(556):eaay6848.3280114510.1126/scitranslmed.aay6848 doi:10.1126/scitranslmed.aay6848 PubMed
375. Grandjean P , Landrigan PJ . Neurobehavioural effects of developmental toxicity. The Lancet Neurology. 2014;13(3):330‐338.2455601010.1016/S1474-4422(13)70278-3PMC4418502 doi:10.1016/S1474-4422(13)70278-3 PubMed
376. Frank CL , Brown JP , Wallace K , Mundy WR , Shafer TJ . From the cover: developmental neurotoxicants disrupt activity in cortical networks on microelectrode arrays: results of screening 86 compounds during neural network formation. Toxicol Sci. 2017;160(1):121‐135.2897355210.1093/toxsci/kfx169 doi:10.1093/toxsci/kfx169 PubMed
377. Liu SH , Bobb JF , Claus Henn B , et al. Bayesian varying coefficient kernel machine regression to assess neurodevelopmental trajectories associated with exposure to complex mixtures. Stat Med. 2018;37(30):4680‐4694.3027758410.1002/sim.7947PMC6522130 doi:10.1002/sim.7947 PubMed
378. Tian S , Sun Y , Shao J , et al. Predicting escitalopram monotherapy response in depression: the role of anterior cingulate cortex. Hum Brain Mapp. 2019;41(5):1249‐1260.3175863410.1002/hbm.24872PMC7268019 doi:10.1002/hbm.24872 PubMed
379. Koutsouleris N , Kahn RS , Chekroud AM , et al. Multisite prediction of 4‐week and 52‐week treatment outcomes in patients with first‐episode psychosis: a machine learning approach. Lancet Psychiatry. 2016;3(10):935‐946.2756952610.1016/S2215-0366(16)30171-7 doi:10.1016/S2215-0366(16)30171-7 PubMed
380. Chekroud AM , Zotti RJ , Shehzad Z , et al. Cross‐trial prediction of treatment outcome in depression: a machine learning approach. Lancet Psychiatry. 2016;3(3):243‐250.2680339710.1016/S2215-0366(15)00471-X doi:10.1016/S2215-0366(15)00471-X PubMed
381. Chekroud AM , Gueorguieva R , Krumholz HM , Trivedi MH , Krystal JH , McCarthy G . Reevaluating the efficacy and predictability of antidepressant treatments: a symptom clustering approach. JAMA Psychiatry. 2017;74(4):370‐378.2824118010.1001/jamapsychiatry.2017.0025PMC5863470 doi:10.1001/jamapsychiatry.2017.0025 PubMed
382. Chekroud AM , Koutsouleris N . The perilous path from publication to practice. Mol Psychiatry. 2018;23(1):24‐25.2911219210.1038/mp.2017.227 doi:10.1038/mp.2017.227 PubMed
383. Chang B , Choi Y , Jeon M , et al. ARPNet: antidepressant response prediction network for major depressive disorder. Genes (Basel). 2019;10:11.10.3390/genes10110907PMC689582931703457 doi:10.3390/genes10110907 PubMed
384. Zhdanov A , Atluri S , Wong W , et al. Use of machine learning for predicting escitalopram treatment outcome from electroencephalography recordings in adult patients with depression. JAMA Netw Open. 2020;3(1):e1918377.3189953010.1001/jamanetworkopen.2019.18377PMC6991244 doi:10.1001/jamanetworkopen.2019.18377 PubMed
385. Wu W , Zhang Y , Jiang J , et al. An electroencephalographic signature predicts antidepressant response in major depression. Nat Biotechnol. 2020;38(4):439‐447.3204216610.1038/s41587-019-0397-3PMC7145761 doi:10.1038/s41587-019-0397-3 PubMed
386. Ichikawa N , Lisi G , Yahata N , et al. Primary functional brain connections associated with melancholic major depressive disorder and modulation by antidepressants. Sci Rep. 2020;10(1):3542.3210308810.1038/s41598-020-60527-zPMC7044159 doi:10.1038/s41598-020-60527-z PubMed
387. Bzdok D , Meyer‐Lindenberg A . Machine learning for precision psychiatry: opportunities and challenges. Biol Psychiatry Cogn Neurosci Neuroimaging. 2018;3(3):223‐230.2948686310.1016/j.bpsc.2017.11.007 doi:10.1016/j.bpsc.2017.11.007 PubMed
388. Dwyer DB , Falkai P , Koutsouleris N . Machine learning approaches for clinical psychology and psychiatry. Annu Rev Clin Psychol. 2018;14(1):91‐118.2940104410.1146/annurev-clinpsy-032816-045037 doi:10.1146/annurev-clinpsy-032816-045037 PubMed
389. Graham S , Depp C , Lee EE , et al. Artificial intelligence for mental health and mental illnesses: an overview. Curr Psychiatry Rep. 2019;21(11):116.3170132010.1007/s11920-019-1094-0PMC7274446 doi:10.1007/s11920-019-1094-0 PubMed
390. Reeve A , Simcox E , Turnbull D . Ageing and Parkinson's disease: why is advancing age the biggest risk factor? Ageing Res Rev. 2014;14:19‐30.2450300410.1016/j.arr.2014.01.004PMC3989046 doi:10.1016/j.arr.2014.01.004 PubMed
391. Glaab E . Computational systems biology approaches for Parkinson's disease. Cell Tissue Res. 2018;373(1):91‐109.2918507310.1007/s00441-017-2734-5PMC6015628 doi:10.1007/s00441-017-2734-5 PubMed
392. Pinto M , Fernandes C , Martins E , et al. Boosting drug discovery for Parkinson's: enhancement of the delivery of a monoamine oxidase‐B inhibitor by brain‐targeted PEGylated polycaprolactone‐based nanoparticles. Pharmaceutics. 2019;11(7):331.10.3390/pharmaceutics11070331PMC668109131336891 doi:10.3390/pharmaceutics11070331 PubMed
393. Taylor JP , Hardy J , Fischbeck KH . Toxic proteins in neurodegenerative disease. Science. 2002;296(5575):1991‐1995.1206582710.1126/science.1067122 doi:10.1126/science.1067122 PubMed
394. Varadi C , Nehez K , Hornyak O , Viskolcz B , Bones J . Serum N‐glycosylation in Parkinson's disease: a novel approach for potential alterations. Molecules (Basel, Switzerland). 2019;24(12):2220.10.3390/molecules24122220PMC663059531200590 doi:10.3390/molecules24122220 PubMed
395. Ishigami N , Tokuda T , Ikegawa M , et al. Cerebrospinal fluid proteomic patterns discriminate Parkinson's disease and multiple system atrophy. Mov Disord. 2012;27(7):851‐857.2267485010.1002/mds.24994 doi:10.1002/mds.24994 PubMed
396. Potashkin JA , Santiago JA , Ravina BM , Watts A , Leontovich AA . Biosignatures for Parkinson's disease and atypical parkinsonian disorders patients. PLOS One. 2012;7(8):e43595.2295271510.1371/journal.pone.0043595PMC3428307 doi:10.1371/journal.pone.0043595 PubMed
397. Abdi F , Quinn JF , Jankovic J , et al. Detection of biomarkers with a multiplex quantitative proteomic platform in cerebrospinal fluid of patients with neurodegenerative disorders. J Alzheimers Dis. 2006;9(3):293‐348.1691484010.3233/jad-2006-9309 doi:10.3233/jad-2006-9309 PubMed
398. Matarazzo M , Arroyo‐Gallego T , Montero P , et al. Remote Monitoring of treatment response in Parkinson's disease: the habit of typing on a computer. Mov Disord. 2019;34(10):1488‐1495.3121146910.1002/mds.27772 doi:10.1002/mds.27772 PubMed
399. Hssayeni MD , Burack MA , Ghoraani B . Automatic assessment of medication states of patients with Parkinson's disease using wearable sensors. Conf Proc IEEE Eng Med Biol Soc. 2016;2016:6082‐6085.10.1109/EMBC.2016.759211628269640 doi:10.1109/EMBC.2016.7592116 PubMed
400. Nguyen V , Kunz H , Taylor P , Acosta D . Insights into pharmacotherapy management for Parkinson's disease patients using wearables activity data. Stud Health Technol Inform. 2018;247:156‐160.29677942 PubMed
401. Hssayeni MD , Burack MA , Jimenez‐Shahed J , Ghoraani B . Assessment of response to medication in individuals with Parkinson's disease. Med Eng Phys. 2019;67:33‐43.3087681710.1016/j.medengphy.2019.03.002 doi:10.1016/j.medengphy.2019.03.002 PubMed
402. Asakawa T , Sugiyama K , Nozaki T , et al. Can the latest computerized technologies revolutionize conventional assessment tools and therapies for a neurological disease? The example of Parkinson's disease. Neurol Med Chir (Tokyo). 2019;59(3):69‐78.3076065710.2176/nmc.ra.2018-0045PMC6434424 doi:10.2176/nmc.ra.2018-0045 PubMed
403. Shao YM , Ma X , Paira P , et al. Discovery of indolylpiperazinylpyrimidines with dual‐target profiles at adenosine A2A and dopamine D2 receptors for Parkinson's disease treatment. PLOS One. 2018;13(1):e0188212.2930411310.1371/journal.pone.0188212PMC5755735 doi:10.1371/journal.pone.0188212 PubMed
404. Sebastian‐Perez V , Martinez MJ , Gil C , Campillo NE , Martinez A , Ponzoni I . QSAR modelling to identify LRRK2 inhibitors for Parkinson's disease. J Integr Bioinform. 2019;16(1):20180063.10.1515/jib-2018-0063PMC679885930763264 doi:10.1515/jib-2018-0063 PubMed
405. Johnston TH , Lacoste AMB , Visanji NP , Lang AE , Fox SH , Brotchie JM . Repurposing drugs to treat l‐DOPA‐induced dyskinesia in Parkinson's disease. Neuropharmacology. 2019;147:11‐27.2990742410.1016/j.neuropharm.2018.05.035 doi:10.1016/j.neuropharm.2018.05.035 PubMed
406. Monzel AS , Hemmer K , Kaoma T , et al. Machine learning‐assisted neurotoxicity prediction in human midbrain organoids. Parkinsonism Rel Disord. 2020;75:105‐109.10.1016/j.parkreldis.2020.05.01132534431 doi:10.1016/j.parkreldis.2020.05.011 PubMed
407. Hughes GL , Lones MA , Bedder M , Currie PD , Smith SL , Pownall ME . Machine learning discriminates a movement disorder in a zebrafish model of Parkinson's disease. Dis Model Mech. 2020;13(10):dmm045815.3285969610.1242/dmm.045815PMC7578351 doi:10.1242/dmm.045815 PubMed
408. Goedert M , Spillantini MG . A century of Alzheimer's disease. Science. 2006;314(5800):777‐781.1708244710.1126/science.1132814 doi:10.1126/science.1132814 PubMed
409. Mangialasche F , Solomon A , Winblad B , Mecocci P , Kivipelto M . Alzheimer's disease: clinical trials and drug development. Lancet Neurol. 2010;9(7):702‐716.2061034610.1016/S1474-4422(10)70119-8 doi:10.1016/S1474-4422(10)70119-8 PubMed
410. Misra S , Medhi B . Drug development status for Alzheimer's disease: present scenario. Neurol Sci. 2013;34(6):831‐839.2339289710.1007/s10072-013-1316-x doi:10.1007/s10072-013-1316-x PubMed
411. Cummings JL , Morstorf T , Zhong K . Alzheimer's disease drug‐development pipeline: few candidates, frequent failures. Alzheimer's Res Ther. 2014;6(4):37.2502475010.1186/alzrt269PMC4095696 doi:10.1186/alzrt269 PubMed
412. Khan S , Barve KH , Kumar MS . Recent advancements in pathogenesis, diagnostics and treatment of Alzheimer's disease. Curr Neuropharmacol. 2020;18:1106‐1125.3248411010.2174/1570159X18666200528142429PMC7709159 doi:10.2174/1570159X18666200528142429 PubMed
413. Louros N , Orlando G , de Vleeschouwer M , Rousseau F , Schymkowitz J . Structure‐based machine‐guided mapping of amyloid sequence space reveals uncharted sequence clusters with higher solubilities. Nat Commun. 2020;11(1):3314.3262086110.1038/s41467-020-17207-3PMC7335209 doi:10.1038/s41467-020-17207-3 PubMed
414. Sügis E , Dauvillier J , Leontjeva A , et al. HENA, heterogeneous network‐based data set for Alzheimer's disease. Sci Data. 2019;6(1):151.3141332510.1038/s41597-019-0152-0PMC6694132 doi:10.1038/s41597-019-0152-0 PubMed
415. Hung TC , Lee WY , Chen KB , Chan YC , Lee CC , Chen CY . In silico investigation of traditional Chinese medicine compounds to inhibit human histone deacetylase 2 for patients with Alzheimer's disease. BioMed Res Int. 2014;2014:769867. 10.1155/2014/769867 PMC409043625045700 doi:10.1155/2014/769867 PubMed
416. Cavas L , Topcam G , Gundogdu‐Hizliates C , Ergun Y . Neural network modeling of AChE inhibition by new carbazole‐bearing oxazolones. Interdiscip Sci. 2019;11(1):95‐107.2923621410.1007/s12539-017-0245-4 doi:10.1007/s12539-017-0245-4 PubMed
417. Lee J , Kumar S , Lee SY , Park SJ , Kim MH . Development of predictive models for identifying potential S100A9 inhibitors based on machine learning methods. Front Chem. 2019;7:779.3182491910.3389/fchem.2019.00779PMC6886474 doi:10.3389/fchem.2019.00779 PubMed
418. Jamal S , Grover A , Grover S . Machine learning from molecular dynamics trajectories to predict caspase‐8 inhibitors against Alzheimer's disease. Front Pharmacol. 2019;10:780.3135449410.3389/fphar.2019.00780PMC6639425 doi:10.3389/fphar.2019.00780 PubMed
419. Miyazaki Y , Ono N , Huang M , Altaf‐Ul‐Amin M , Kanaya S . Comprehensive exploration of target‐specific ligands using a graph convolution neural network. Mol Inf. 2020;39(1–2):e1900095.10.1002/minf.201900095PMC705050431815371 doi:10.1002/minf.201900095 PubMed
420. Kleandrova VV , Speck‐Planche A . PTML modeling for Alzheimer's disease: design and prediction of virtual multi‐target inhibitors of GSK3B, HDAC1, and HDAC6. Curr Top Med Chem. 2020;20(19):1661‐1676.3251531110.2174/1568026620666200607190951 doi:10.2174/1568026620666200607190951 PubMed
421. Riccardo C , Michael G‐D , Maria Natália DSC . Developing a multi‐target model to predict the activity of monoamine oxidase A and B drugs. Curr Top Med Chem. 2020;20(18):1593‐1600.3249319310.2174/1568026620666200603121224 doi:10.2174/1568026620666200603121224 PubMed
422. Gupta R , Ambasta RK , Kumar P . Identification of novel class I and class IIb histone deacetylase inhibitor for Alzheimer's disease therapeutics. Life Sci. 2020;256:117912.3250475510.1016/j.lfs.2020.117912 doi:10.1016/j.lfs.2020.117912 PubMed
423. Fang J , Li Y , Liu R , et al. Discovery of multitarget‐directed ligands against Alzheimer's disease through systematic prediction of chemical‐protein interactions. J Chem Inf Model. 2015;55(1):149‐164.2553179210.1021/ci500574n doi:10.1021/ci500574n PubMed
424. Fang J , Wang L , Li Y , et al. AlzhCPI: a knowledge base for predicting chemical‐protein interactions towards Alzheimer's disease. PLOS One. 2017;12(5):e0178347.2854250510.1371/journal.pone.0178347PMC5460905 doi:10.1371/journal.pone.0178347 PubMed
425. Pang XC , Kang D , Fang JS , et al. Network pharmacology‐based analysis of Chinese herbal Naodesheng formula for application to Alzheimer's disease. Chin J Nat Med. 2018;16(1):53‐62.2942559010.1016/S1875-5364(18)30029-3 doi:10.1016/S1875-5364(18)30029-3 PubMed
426. Grisoni F , Merk D , Friedrich L , Schneider G . Design of natural‐product‐inspired multitarget ligands by machine learning. ChemMedChem. 2019;14(12):1129‐1134.3097367210.1002/cmdc.201900097 doi:10.1002/cmdc.201900097 PubMed
427. Thompson CA . FDA approves galantamine for Alzheimer's disease. Am J Health Syst Pharm. 2001;58(8):649.10.1093/ajhp/58.8.649a11329755 doi:10.1093/ajhp/58.8.649a PubMed
428. Jamal S , Goyal S , Shanker A , Grover A . Integrating network, sequence and functional features using machine learning approaches towards identification of novel Alzheimer genes. BMC Genomics. 2016;17(1):807.2775622310.1186/s12864-016-3108-1PMC5070370 doi:10.1186/s12864-016-3108-1 PubMed
429. Hampel H , Williams C , Etcheto A , et al. A precision medicine framework using artificial intelligence for the identification and confirmation of genomic biomarkers of response to an Alzheimer's disease therapy: analysis of the blarcamesine (ANAVEX2‐73) Phase 2a clinical study. Alzheimers Dement (NY). 2020;6(1):e12013.10.1002/trc2.12013PMC716737432318621 doi:10.1002/trc2.12013 PubMed
430. Lu H , Zhang J , Liang Y , et al. Network topology and machine learning analyses reveal microstructural white matter changes underlying Chinese medicine Dengzhan Shengmai treatment on patients with vascular cognitive impairment. Pharmacol Res. 2020;156:104773.3224402810.1016/j.phrs.2020.104773 doi:10.1016/j.phrs.2020.104773 PubMed
431. Hemmerling TM , Taddei R , Wehbe M , Morse J , Cyr S , Zaouter C . Robotic anesthesia—a vision for the future of anesthesia. Transl Med UniSa. 2011;1:1‐20.23905028PMC3728848 PubMed
432. Zaouter C , Joosten A , Rinehart J , Struys MMRF , Hemmerling TM . Autonomous systems in anesthesia: where do we stand in 2020? A narrative review. Anesth Analg. 2020;130(5):1120‐1132.3228712010.1213/ANE.0000000000004646 doi:10.1213/ANE.0000000000004646 PubMed
433. Char DS , Burgart A . Machine‐learning implementation in clinical anesthesia: opportunities and challenges. Anesth Analg. 2020;130(6):1709‐1712.3192299610.1213/ANE.0000000000004656PMC7654525 doi:10.1213/ANE.0000000000004656 PubMed
434. Ramaswamy SM , Weerink MAS , Struys MMRF , Nagaraj SB . Dexmedetomidine‐induced deep sedation mimics non‐rapid eye movement stage 3 sleep: large‐scale validation using machine learning. Sleep. 2020. 10.1093/sleep/zsaa167 PMC787942032860500 doi:10.1093/sleep/zsaa167 PubMed
435. Kashkooli K , Polk SL , Hahm EY , et al. Improved tracking of sevoflurane anesthetic states with drug‐specific machine learning models. J Neural Eng. 2020;17(4):046020.3248568510.1088/1741-2552/ab98daPMC7540939 doi:10.1088/1741-2552/ab98da PubMed
436. Belur Nagaraj S , Ramaswamy SM , Weerink MAS , Struys M . Predicting deep hypnotic state from sleep brain rhythms using deep learning: a data‐repurposing approach. Anesth Analg. 2020;130(5):1211‐1221.3228712810.1213/ANE.0000000000004651PMC7147424 doi:10.1213/ANE.0000000000004651 PubMed
437. Kang AR , Lee J , Jung W , et al. Development of a prediction model for hypotension after induction of anesthesia using machine learning. PLOS One. 2020;15(4):e0231172.3229829210.1371/journal.pone.0231172PMC7162491 doi:10.1371/journal.pone.0231172 PubMed
438. Ermer SC , Farney RJ , Johnson KB , Orr JA , Egan TD , Brewer LM . An automated algorithm incorporating poincare analysis can quantify the severity of opioid‐induced ataxic breathing. Anesth Analg. 2020;130(5):1147‐1156.3228712210.1213/ANE.0000000000004498 doi:10.1213/ANE.0000000000004498 PubMed
439. Moghadam MC , Abad EMK , Bagherzadeh N , Ramsingh D , Li GP , Kain ZN . A machine‐learning approach to predicting hypotensive events in ICU settings. Comput Biol Med. 2020;118:103626.3217432810.1016/j.compbiomed.2020.103626 doi:10.1016/j.compbiomed.2020.103626 PubMed
440. Wang R , Wang S , Duan N , Wang Q . From patient‐controlled analgesia to artificial intelligence‐assisted patient‐controlled analgesia: practices and perspectives. Front Med. 2020;7(145).10.3389/fmed.2020.00145PMC732606432671076 doi:10.3389/fmed.2020.00145 PubMed
441. Johnson A , Yang F , Gollarahalli S , et al. Use of mobile health apps and wearable technology to assess changes and predict pain during treatment of acute pain in sickle cell disease: feasibility study. JMIR Mhealth Uhealth. 2019;7(12):e13671.3178959910.2196/13671PMC6915456 doi:10.2196/13671 PubMed
442. Nair AA , Velagapudi MA , Lang JA , et al. Machine learning approach to predict postoperative opioid requirements in ambulatory surgery patients. PLOS One. 2020;15(7):e0236833.3273560410.1371/journal.pone.0236833PMC7394436 doi:10.1371/journal.pone.0236833 PubMed
443. Walter S , Al‐Hamadi A , Gruss S , Frisch S , Traue HC , Werner P . Multimodal recognition of pain intensity and pain modality with machine learning. Schmerz. 2020;34(5):400‐409.3229158810.1007/s00482-020-00468-8 doi:10.1007/s00482-020-00468-8 PubMed
444. Gudin J , Mavroudi S , Korfiati A , Theofilatos K , Dietze D , Hurwitz P . Reducing opioid prescriptions by identifying responders on topical analgesic treatment using an individualized medicine and predictive analytics approach. J Pain Res. 2020;13:1255‐1266.3254718610.2147/JPR.S246503PMC7266406 doi:10.2147/JPR.S246503 PubMed
445. Lopez‐Martinez D , Eschenfeldt P , Ostvar S , Ingram M , Hur C , Picard R . Deep reinforcement learning for optimal critical care pain management with morphine using dueling double‐deep Q networks. Annu Int Conf IEEE Eng Med Biol Soc. 2019;2019:3960‐3963.3194673910.1109/EMBC.2019.8857295 doi:10.1109/EMBC.2019.8857295 PubMed
446. Parthipan A , Banerjee I , Humphreys K , et al. Predicting inadequate postoperative pain management in depressed patients: a machine learning approach. PLOS One. 2019;14(2):e0210575.3072623710.1371/journal.pone.0210575PMC6364959 doi:10.1371/journal.pone.0210575 PubMed
447. Wei M , Liao Y , Liu J , et al. EEG beta‐band spectral entropy can predict the effect of drug treatment on pain in patients with herpes zoster [published online ahead of print July 14, 2020]. J Clin Neurophysiol. 2020. 10.1097/WNP.0000000000000758 32675727 doi:10.1097/WNP.0000000000000758 PubMed
448. Ferroni P , Zanzotto FM , Scarpato N , et al. Machine learning approach to predict medication overuse in migraine patients. Comput Struct Biotechnol J. 2020;18:1487‐1496.3263704610.1016/j.csbj.2020.06.006PMC7327028 doi:10.1016/j.csbj.2020.06.006 PubMed
449. Lee S , Wei S , White V , Bain P , Baker C , Li J . Classification of opioid usage through semi‐supervised learning for total joint replacement patients. IEEE J Biomed Health Inform, 2020:1.10.1109/JBHI.2020.299297332386170 doi:10.1109/JBHI.2020.2992973 PubMed
450. Zhang Y , Fatemi P , Medress Z , et al. A predictive‐modeling‐based screening tool for prolonged opioid use after surgical management of low back and lower extremity pain. Spine J. 2020;20(8):1184‐1195.3244580310.1016/j.spinee.2020.05.098 doi:10.1016/j.spinee.2020.05.098 PubMed
451. Chidambaran V , Ashton M , Martin LJ , Jegga AG . Systems biology‐based approaches to summarize and identify novel genes and pathways associated with acute and chronic postsurgical pain. J Clin Anesth. 2020;62:109738.3205825910.1016/j.jclinane.2020.109738PMC7276001 doi:10.1016/j.jclinane.2020.109738 PubMed
452. Kong W , Tu X , Huang W , Yang Y , Xie Z , Huang Z . Prediction and optimization of NaV1.7 sodium channel inhibitors based on machine learning and simulated annealing. J Chem Inf Model. 2020;60(6):2739‐2753.3242131810.1021/acs.jcim.9b01180 doi:10.1021/acs.jcim.9b01180 PubMed
453. Feng Z , Chen M , Shen M , Liang T , Chen H , Xie XQ . Pain‐CKB, a pain‐domain‐specific chemogenomics knowledgebase for target identification and systems pharmacology research. J Chem Inf Model. 2020;60:4429‐4435.3278669410.1021/acs.jcim.0c00633 doi:10.1021/acs.jcim.0c00633 PubMed
454. Startups using artificial intelligence in drug discovery. https://blog.benchsci.com/startups-using-artificial-intelligence-in-drug-discovery
455. Chan HCS , Shan H , Dahoun T , Vogel H , Yuan S . Advancing drug discovery via artificial intelligence. Trends Pharmacol Sci. 2019;40(8):592‐604.3132011710.1016/j.tips.2019.06.004 doi:10.1016/j.tips.2019.06.004 PubMed
456. Mak K‐K , Pichika MR . Artificial intelligence in drug development: present status and future prospects. Drug Discovery Today. 2019;24(3):773‐780.3047242910.1016/j.drudis.2018.11.014 doi:10.1016/j.drudis.2018.11.014 PubMed
457. Hinton GE , Krizhevsky A , Wang SD . Transforming Auto‐Encoders. Berlin, Heidelberg, Germany: Springer; 2011.
458. Becker G , Bolbos R , Costes N , Redoute J , Newman‐Tancredi A , Zimmer L . Selective serotonin 5‐HT1A receptor biased agonists elicit distinct brain activation patterns: a pharmacoMRI study. Sci Rep. 2016;6:26633.2721107810.1038/srep26633PMC4876409 doi:10.1038/srep26633 PubMed
459. Downey D , Dutta A , McKie S , et al. Comparing the actions of lanicemine and ketamine in depression: key role of the anterior cingulate. Eur Neuropsychopharmacol. 2016;26(6):994‐1003.2713302910.1016/j.euroneuro.2016.03.006 doi:10.1016/j.euroneuro.2016.03.006 PubMed
460. von Pföstl V , Li J , Zaldivar D , et al. Effects of lactate on the early visual cortex of non‐human primates, investigated by pharmaco‐MRI and neurochemical analysis. Neuroimage. 2012;61(1):98‐105.2242635010.1016/j.neuroimage.2012.02.082 doi:10.1016/j.neuroimage.2012.02.082 PubMed
461. Carmichael O , Schwarz AJ , Chatham CH , et al. The role of fMRI in drug development. Drug Discovery Today. 2018;23(2):333‐348.2915475810.1016/j.drudis.2017.11.012PMC5931333 doi:10.1016/j.drudis.2017.11.012 PubMed
462. Urban A , Golgher L , Brunner C , et al. Understanding the neurovascular unit at multiple scales: advantages and limitations of multi‐photon and functional ultrasound imaging. Adv Drug Deliv Rev. 2017;119:73‐100.2877871410.1016/j.addr.2017.07.018 doi:10.1016/j.addr.2017.07.018 PubMed
463. Macé E , Montaldo G , Cohen I , Baulac M , Fink M , Tanter M . Functional ultrasound imaging of the brain. Nat Methods. 2011;8(8):662‐664.2172530010.1038/nmeth.1641 doi:10.1038/nmeth.1641 PubMed
464. Vidal B , Droguerre M , Valdebenito M , et al. Pharmaco‐fUS for characterizing drugs for Alzheimer's disease—the case of THN201, a drug combination of donepezil plus mefloquine. Front Neurosci. 2020;14:835.3290347010.3389/fnins.2020.00835PMC7437134 doi:10.3389/fnins.2020.00835 PubMed
465. Vidal B , Droguerre M , Venet L , et al. Functional ultrasound imaging to study brain dynamics: application of pharmaco‐fUS to atomoxetine. Neuropharmacology. 2020;179:108273.3280102510.1016/j.neuropharm.2020.108273 doi:10.1016/j.neuropharm.2020.108273 PubMed
466. Rabut C , Ferrier J , Bertolo A , et al. Pharmaco‐fUS: quantification of pharmacologically‐induced dynamic changes in brain perfusion and connectivity by functional ultrasound imaging in awake mice. Neuroimage. 2020;222:117231.3279565910.1016/j.neuroimage.2020.117231 doi:10.1016/j.neuroimage.2020.117231 PubMed
467. Chuang KV , Keiser MJ . Adversarial controls for scientific machine learning. ACS Chem Biol. 2018;13(10):2819‐2821.3033667010.1021/acschembio.8b00881 doi:10.1021/acschembio.8b00881 PubMed
468. Schneider G . Mind and machine in drug design. Nat Mach Intell. 2019;1(3):128‐130.