Discovery of drug mode of action and drug repositioning from transcriptional responses

Recycled Translation: Repurposing Drugs for Stroke

Stroke, which continues to be a leading cause of death and long-term disability worldwide, has often been described as a clinical graveyard. While multiple small molecule therapeutics have undergone clinical trials in stroke, currently only one Food and Drug Administration (FDA)-approved medication exists for the treatment of stroke, the biological, recombinant tissue plasminogen activator (rt-PA). Repurposing of therapeutics which have previously gained FDA approval for alternative indications serves as a prospective option for stroke therapeutic translation. In contrast to de novo drug development, repurposing strategies have patient-centered and economic advantages. These include increased safety, increased chance of approval, decreased time to approval, and decreased capital investment. Presently, 37 active stroke clinical trials utilize repurposed therapeutics with various initial indications and dosing paradigms. The currently studied repurposed therapeutics fall into six mechanistic categories: (1) anticoagulation; (2) vasculature integrity, response, or red blood cell (RBC) alterations; (3) immune system regulation; (4) neurotransmission; and (5) neuroprotection. Directed hypothesis-driven computational investigation utilizing drug databases, in silico drug-protein interaction modeling, genomic data, and consensus methodology can determine if the current mechanistic repurposing categories have the highest chance of translational success or if other mechanistic avenues should be explored. With this increased focus on repurposed therapeutic strategies over de novo strategies, evolution and optimization of regulatory protections are needed to incentivize innovators and investigators.

Estimating Metastatic Risk of Pancreatic Ductal Adenocarcinoma at Single-Cell Resolution

Pancreatic ductal adenocarcinoma (PDAC) is characterized by intra-tumoral heterogeneity, and patients are always diagnosed after metastasis. Thus, finding out how to effectively estimate metastatic risk underlying PDAC is necessary. In this study, we proposed scMetR to evaluate the metastatic risk of tumor cells based on single-cell RNA sequencing (scRNA-seq) data. First, we identified diverse cell types, including tumor cells and other cell types. Next, we grouped tumor cells into three sub-populations according to scMetR score, including metastasis-featuring tumor cells (MFTC), transitional metastatic tumor cells (TransMTC), and conventional tumor cells (ConvTC). We identified metastatic signature genes (MSGs) through comparing MFTC and ConvTC. Functional enrichment analysis showed that up-regulated MSGs were enriched in multiple metastasis-associated pathways. We also found that patients with high expression of up-regulated MSGs had worse prognosis. Spatial mapping of MFTC showed that they are preferentially located in the cancer and duct epithelium region, which was enriched with the ductal cells' associated inflammation. Further, we inferred cell-cell interactions, and observed that interactions of the ADGRE5 signaling pathway, which is associated with metastasis, were increased in MFTC compared to other tumor sub-populations. Finally, we predicted 12 candidate drugs that had the potential to reverse expression of MSGs. Taken together, we have proposed scMetR to estimate metastatic risk in PDAC patients at single-cell resolution which might facilitate the dissection of tumor heterogeneity.

Deep learning in drug discovery: an integrative review and future challenges

Recently, using artificial intelligence (AI) in drug discovery has received much attention since it significantly shortens the time and cost of developing new drugs. Deep learning (DL)-based approaches are increasingly being used in all stages of drug development as DL technology advances, and drug-related data grows. Therefore, this paper presents a systematic Literature review (SLR) that integrates the recent DL technologies and applications in drug discovery Including, drug-target interactions (DTIs), drug-drug similarity interactions (DDIs), drug sensitivity and responsiveness, and drug-side effect predictions. We present a review of more than 300 articles between 2000 and 2022. The benchmark data sets, the databases, and the evaluation measures are also presented. In addition, this paper provides an overview of how explainable AI (XAI) supports drug discovery problems. The drug dosing optimization and success stories are discussed as well. Finally, digital twining (DT) and open issues are suggested as future research challenges for drug discovery problems. Challenges to be addressed, future research directions are identified, and an extensive bibliography is also included.

Regulome-based characterization of drug activity across the human diseasome

Drugs are expected to recover the cell system away from the impaired state to normalcy through disease treatment. However, the understanding of gene regulatory machinery underlying drug activity or disease pathogenesis is far from complete. Here, we perform large-scale regulome analysis for various diseases in terms of gene regulatory machinery. Transcriptome signatures were converted into regulome signatures of transcription factors by integrating publicly available ChIP-seq data. Regulome-based correlations between diseases and their approved drugs were much clearer than the transcriptome-based correlations. For example, an inverse correlation was observed for cancers, whereas a positive correlation was observed for immune system diseases. After demonstrating the usefulness of the regulome-based drug discovery method in terms of accuracy and applicability, we predicted new drugs for nonsmall cell lung cancer and validated the anticancer activity in vitro. The proposed method is useful for understanding disease–disease relationships and drug discovery.

DrSim: Similarity Learning for Transcriptional Phenotypic Drug Discovery

Transcriptional phenotypic drug discovery has achieved great success, and various compound perturbation-based data resources, such as connectivity map (CMap) and library of integrated network-based cellular signatures (LINCS), have been presented. Computational strategies fully mining these resources for phenotypic drug discovery have been proposed. Among them, the fundamental issue is to define the proper similarity between transcriptional profiles. Traditionally, such similarity has been defined in an unsupervised way. However, due to the high dimensionality and the existence of high noise in high-throughput data, similarity defined in the traditional way lacks robustness and has limited performance. To this end, we present DrSim, which is a learning-based framework that automatically infers similarity rather than defining it. We evaluated DrSim on publicly available in vitro and in vivo datasets in drug annotation and repositioning. The results indicated that DrSim outperforms the existing methods. In conclusion, by learning transcriptional similarity, DrSim facilitates the broad utility of high-throughput transcriptional perturbation data for phenotypic drug discovery. The source code and manual of DrSim are available at https://github.com/bm2-lab/DrSim/.

Recent Advances of Utilizing Artificial Intelligence in Lab on a Chip for Diagnosis and Treatment

Nowadays, artificial intelligence (AI) creates numerous promising opportunities in the life sciences. AI methods can be significantly advantageous for analyzing the massive datasets provided by biotechnology systems for biological and biomedical applications. Microfluidics, with the developments in controlled reaction chambers, high-throughput arrays, and positioning systems, generate big data that is not necessarily analyzed successfully. Integrating AI and microfluidics can pave the way for both experimental and analytical throughputs in biotechnology research. Microfluidics enhances the experimental methods and reduces the cost and scale, while AI methods significantly improve the analysis of huge datasets obtained from high-throughput and multiplexed microfluidics. This review briefly presents a survey of the role of AI and microfluidics in biotechnology. Also, the incorporation of AI with microfluidics is comprehensively investigated. Specifically, recent studies that perform flow cytometry cell classification, cell isolation, and a combination of them by gaining from both AI methods and microfluidic techniques are covered. Despite all current challenges, various fields of biotechnology can be remarkably affected by the combination of AI and microfluidic technologies. Some of these fields include point-of-care systems, precision, personalized medicine, regenerative medicine, prognostics, diagnostics, and treatment of oncology and non-oncology-related diseases.

A Survey on Computational Methods for Investigation on ncRNA-Disease Association through the Mode of Action Perspective

Molecular and sequencing technologies have been successfully used in decoding biological mechanisms of various diseases. As revealed by many novel discoveries, the role of non-coding RNAs (ncRNAs) in understanding disease mechanisms is becoming increasingly important. Since ncRNAs primarily act as regulators of transcription, associating ncRNAs with diseases involves multiple inference steps. Leveraging the fast-accumulating high-throughput screening results, a number of computational models predicting ncRNA-disease associations have been developed. These tools suggest novel disease-related biomarkers or therapeutic targetable ncRNAs, contributing to the realization of precision medicine. In this survey, we first introduce the biological roles of different ncRNAs and summarize the databases containing ncRNA-disease associations. Then, we suggest a new trend in recent computational prediction of ncRNA-disease association, which is the mode of action (MoA) network perspective. This perspective includes integrating ncRNAs with mRNA, pathway and phenotype information. In the next section, we describe computational methodologies widely used in this research domain. Existing computational studies are then summarized in terms of their coverage of the MoA network. Lastly, we discuss the potential applications and future roles of the MoA network in terms of integrating biological mechanisms for ncRNA-disease associations.

Literature-based drug-drug similarity for drug repurposing: impact of Medical Subject Headings term refinement and hierarchical clustering

Background: We describe herein, an improved procedure for drug repurposing based on refined Medical Subject Headings (MeSH) terms and hierarchical clustering method. Materials & methods: In the present study, we have employed MeSH data from MEDLINE (2019), 1669 US FDA approved drugs from Open FDA and a refined set of MeSH terms. Refinement of MeSH terms was performed to include terms related to mechanistic information of drugs and diseases. Results and Conclusions: In-depth analysis of the results obtained, demonstrated greater efficiency of the proposed approach, based on refined MeSH terms and hierarchical clustering, in terms of number of selected drug candidates for repurposing. Further, analysis of misclustering and size of noise clusters suggest that the proposed approach is reliable and can be employed in drug repurposing.

Single-cell transcriptional pharmacodynamics of trifluridine in a tumor-immune model

Understanding the immunological effects of chemotherapy is of great importance, especially now that we have entered an era where ever-increasing pre-clinical and clinical efforts are put into combining chemotherapy and immunotherapy to combat cancer. Single-cell RNA sequencing (scRNA-seq) has proved to be a powerful technique with a broad range of applications, studies evaluating drug effects in co-cultures of tumor and immune cells are however scarce. We treated a co-culture comprised of human colorectal cancer (CRC) cells and peripheral blood mononuclear cells (PBMCs) with the nucleoside analogue trifluridine (FTD) and used scRNA-seq to analyze posttreatment gene expression profiles in thousands of individual cancer and immune cells concurrently. ScRNA-seq recapitulated major mechanisms of action previously described for FTD and provided new insight into possible treatment-induced effects on T-cell mediated antitumor responses.

Single-cell analysis highlights differences in druggable pathways underlying adaptive or fibrotic kidney regeneration

The kidney has tremendous capacity to repair after acute injury, however, pathways guiding adaptive and fibrotic repair are poorly understood. We developed a model of adaptive and fibrotic kidney regeneration by titrating ischemic injury dose. We performed detailed biochemical and histological analysis and profiled transcriptomic changes at bulk and single-cell level (> 110,000 cells) over time. Our analysis highlights kidney proximal tubule cells as key susceptible cells to injury. Adaptive proximal tubule repair correlated with fatty acid oxidation and oxidative phosphorylation. We identify a specific maladaptive/profibrotic proximal tubule cluster after long ischemia, which expresses proinflammatory and profibrotic cytokines and myeloid cell chemotactic factors. Druggability analysis highlights pyroptosis/ferroptosis as vulnerable pathways in these profibrotic cells. Pharmacological targeting of pyroptosis/ferroptosis in vivo pushed cells towards adaptive repair and ameliorates fibrosis. In summary, our single-cell analysis defines key differences in adaptive and fibrotic repair and identifies druggable pathways for pharmacological intervention to prevent kidney fibrosis.

Prediction of the Mechanism of Sodium Butyrate against Radiation-Induced Lung Injury in Non-Small Cell Lung Cancer Based on Network Pharmacology and Molecular Dynamic Simulations and Molecular Dynamic Simulations

Background

Radiation-induced lung injury (RILI) is a severe side effect of radiotherapy for non-small cell lung cancer (NSCLC) ,and one of the major hindrances to improve the efficacy of radiotherapy. Previous studies have confirmed that sodium butyrate (NaB) has potential of anti-radiation toxicity. However, the mechanism of the protective effect of NaB against RILI has not yet been clarified. This study aimed to explore the underlying protective mechanisms of NaB against RILI in NSCLC through network pharmacology, molecular docking, molecular dynamic simulations and in vivo experiments.

Methods

The predictive target genes of NaB were obtained from the PharmMapper database and the literature review. The involved genes of RILI and NSCLC were predicted using OMIM and GeneCards database. The intersectional genes of drug and disease were identified using the Venny tool and uploaded to the Cytoscape software to identify 5 core target genes of NaB associated with RILI. The correlations between the 5 core target genes and EGFR, PD-L1, immune infiltrates, chemokines and chemokine receptors were analyzed using TIMER 2.0, TIMER and TISIDB databases. We constructed the mechanism maps of the 3 key signaling pathways using the KEGG database based on the results of GO and KEGG analyses from Metascape database. The 5 core target genes and drug were docked using the AutoDock Vina tool and visualized using PyMOL software. GROMACS software was used to perform 100 ns molecular dynamics simulation. Irradiation-induced lung injury model in mice were established to assess the therapeutic effects of NaB.

Results

A total of 51 intersectional genes involved in NaB against RILI in NSCLC were identified. The 5 core target genes were AKT1, TP53, NOTCH1, SIRT1, and PTEN. The expressions of the 5 core target genes were significantly associated with EGFR, PD-L1, immune infiltrates, chemokines and chemokine receptors, respectively. The results from GO analysis of the 51 intersectional genes revealed that the biological processes were focused on the regulation of smooth muscle cell proliferation, oxidative stress and cell death, while the three key KEGG pathways were enriched in PI3K-Akt signal pathway, p53 signal pathway, and FOXO signal pathway. The docking of NaB with the 5 core target genes showed affinity and stability, especially AKT1. In vivo experiments showed that NaB treatment significantly protected mice from RILI, with reduced lung histological damage. In addition, NaB treatment significantly inhibited the PI3K/Akt signaling pathway.

Conclusions

NaB may protect patients from RILI in NSCLC through multiple target genes including AKT1, TP53, NOTCH1, SIRT1 and PTEN, with multiple signaling pathways involving, including PI3K-Akt pathway, p53 pathway, and FOXO pathways. Our findings effectively provide a feasible theoretical basis to further elucidate the mechanism of NaB in the treatment of RILI.

A comprehensive review of Artificial Intelligence and Network based approaches to drug repurposing in Covid-19

Conventional drug discovery and development is tedious and time-taking process; because of which it has failed to keep the required pace to mitigate threats and cater demands of viral and re-occurring diseases, such as Covid-19. The main reasons of this delay in traditional drug development are: high attrition rates, extensive time requirements, and huge financial investment with significant risk. The effective solution to de novo drug discovery is drug repurposing. Previous studies have shown that the network-based approaches and analysis are versatile platform for repurposing as the network biology is used to model the interactions between variety of biological concepts. Herein, we provide a comprehensive background of machine learning and deep learning in drug repurposing while specifically focusing on the applications of network-based approach to drug repurposing in Covid-19, data sources, and tools used. Furthermore, use of network proximity, network diffusion, and AI on network-based drug repurposing for Covid-19 is well-explained. Finally, limitations of network-based approaches in general and specific to network are stated along with future recommendations for better network-based models.

Computational drug repositioning using similarity constrained weight regularization matrix factorization: A case of COVID-19

Amid the COVID‐19 crisis, we put sizeable efforts to collect a high number of experimentally validated drug–virus association entries from literature by text mining and built a human drug–virus association database. To the best of our knowledge, it is the largest publicly available drug–virus database so far. Next, we develop a novel weight regularization matrix factorization approach, termed WRMF, for in silico drug repurposing by integrating three networks: the known drug–virus association network, the drug–drug chemical structure similarity network, and the virus–virus genomic sequencing similarity network. Specifically, WRMF adds a weight to each training sample for reducing the influence of negative samples (i.e. the drug–virus association is unassociated). A comparison on the curated drug–virus database shows that WRMF performs better than a few state‐of‐the‐art methods. In addition, we selected the other two different public datasets (i.e. Cdataset and HMDD V2.0) to assess WRMF's performance. The case study also demonstrated the accuracy and reliability of WRMF to infer potential drugs for the novel virus. In summary, we offer a useful tool including a novel drug–virus association database and a powerful method WRMF to repurpose potential drugs for new viruses.

Supervised graph co-contrastive learning for drug-target interaction prediction

MOTIVATION

Identification of Drug-Target Interactions (DTIs) is an essential step in drug discovery and repositioning. DTI prediction based on biological experiments is time-consuming and expensive. In recent years, graph learning-based methods have aroused widespread interest and shown certain advantages on this task, where the DTI prediction is often modeled as a binary classification problem of the nodes composed of drug and protein pairs (DPPs). Nevertheless, in many real applications, labeled data are very limited and expensive to obtain. With only a few thousand labeled data, models could hardly recognize comprehensive patterns of DPP node representations, and are unable to capture enough commonsense knowledge, which is required in DTI prediction. Supervised contrastive learning gives an aligned representation of DPP node representations with the same class label. In embedding space, DPP node representations with the same label are pulled together, and those with different labels are pushed apart.

RESULTS

We propose an end-to-end supervised graph co-contrastive learning model for DTI prediction directly from heterogeneous networks. By contrasting the topology structures and semantic features of the drug-protein-pair network, as well as the new selection strategy of positive and negative samples, SGCL-DTI generates a contrastive loss to guide the model optimization in a supervised manner. Comprehensive experiments on three public datasets demonstrate that our model outperforms the SOTA methods significantly on the task of DTI prediction, especially in the case of cold start. Furthermore, SGCL-DTI provides a new research perspective of contrastive learning for DTI prediction.

AVAILABILITY AND IMPLEMENTATION

The research shows that this method has certain applicability in the discovery of drugs, the identification of drug-target pairs and so on.

multi-type neighbors enhanced global topology and pairwise attribute learning for drug-protein interaction prediction

MOTIVATION

Accurate identification of proteins interacted with drugs helps reduce the time and cost of drug development. Most of previous methods focused on integrating multisource data about drugs and proteins for predicting drug-target interactions (DTIs). There are both similarity connection and interaction connection between two drugs, and these connections reflect their relationships from different perspectives. Similarly, two proteins have various connections from multiple perspectives. However, most of previous methods failed to deeply integrate these connections. In addition, multiple drug-protein heterogeneous networks can be constructed based on multiple kinds of connections. The diverse topological structures of these networks are still not exploited completely.

RESULTS

We propose a novel model to extract and integrate multi-type neighbor topology information, diverse similarities and interactions related to drugs and proteins. Firstly, multiple drug-protein heterogeneous networks are constructed according to multiple kinds of connections among drugs and those among proteins. The multi-type neighbor node sequences of a drug node (or a protein node) are formed by random walks on each network and they reflect the hidden neighbor topological structure of the node. Secondly, a module based on graph neural network (GNN) is proposed to learn the multi-type neighbor topologies of each node. We propose attention mechanisms at neighbor node level and at neighbor type level to learn more informative neighbor nodes and neighbor types. A network-level attention is also designed to enhance the context dependency among multiple neighbor topologies of a pair of drug and protein nodes. Finally, the attribute embedding of the drug-protein pair is formulated by a proposed embedding strategy, and the embedding covers the similarities and interactions about the pair. A module based on three-dimensional convolutional neural networks (CNN) is constructed to deeply integrate pairwise attributes. Extensive experiments have been performed and the results indicate GCDTI outperforms several state-of-the-art prediction methods. The recall rate estimation over the top-ranked candidates and case studies on 5 drugs further demonstrate GCDTI's ability in discovering potential drug-protein interactions.

Deep Learning Applications for the accurate identification of low-transcriptional activity drugs and their mechanism of actions

Analysis of drug-induced expression profiles facilitated comprehensive understanding of drug properties. However, many compounds exhibit weak transcription responses though they mostly possess definite pharmacological effects. Actually, as a representative example, over 66.4% of 312,438 molecular signatures in the Library of Integrated Cellular Signatures (LINCS) database exhibit low-transcriptional activities (i.e. TAS-low signatures). When computing the association between TAS-low signatures with shared mechanism of actions (MOAs), commonly used algorithms showed inadequate performance with an average area under receiver operating characteristic curve (AUROC) of 0.55, but the computation accuracy of the same task can be improved by our developed tool Genetic profile activity relationship (GPAR) with an average AUROC of 0.68. Up to 36 out of 74 TAS-low MOAs were well trained with AUROC≥0.7 by GPAR, higher than those by other approaches. Further studies showed that GPAR benefited from the size of training samples more significantly than other approaches. Lastly, in biological validation of the MOA prediction for a TAS-low drug Tropisetron, we found an unreported mechanism that Tropisetron can bind to the glucocorticoid receptor. This study indicated that GPAR can serve as an effective approach for the accurate identification of low-transcriptional activity drugs and their MOAs, thus providing a good tool for drug repurposing with both TAS-low and TAS-high signatures.

Discovery of Novel Drug Candidates for Alzheimer’s Disease by Molecular Network Modeling

To identify the molecular mechanisms and novel therapeutic agents of late-onset Alzheimer’s disease (AD), we performed integrative network analysis using multiple transcriptomic profiles of human brains. With the hypothesis that AD pathology involves the whole cerebrum, we first identified co-expressed modules across multiple cerebral regions of the aging human brain. Among them, two modules (M3 and M8) consisting of 1,429 protein-coding genes were significantly enriched with AD-correlated genes. Differential expression analysis of microarray, bulk RNA-sequencing (RNA-seq) data revealed the dysregulation of M3 and M8 across different cerebral regions in both normal aging and AD. The cell-type enrichment analysis and differential expression analysis at the single-cell resolution indicated the extensive neuronal vulnerability in AD pathogenesis. Transcriptomic-based drug screening from Connectivity Map proposed Gly-His-Lys acetate salt (GHK) as a potential drug candidate that could probably restore the dysregulated genes of the M3 and M8 network. Pretreatment with GHK showed a neuroprotective effect against amyloid-beta-induced injury in differentiated human neuron-like SH-SY5Y cells. Taken together, our findings uncover a dysregulated network disrupted across multiple cerebral regions in AD and propose pretreatment with GHK as a novel neuroprotective strategy against AD.

Characterization of Altered Molecular Pathways in the Entorhinal Cortex of Alzheimer’s Disease Patients and In Silico Prediction of Potential Repurposable Drugs

Alzheimer’s disease (AD) is the most common cause of dementia worldwide and is characterized by a progressive decline in cognitive functions. Accumulation of amyloid-β plaques and neurofibrillary tangles are a typical feature of AD neuropathological changes. The entorhinal cortex (EC) is the first brain area associated with pathologic changes in AD, even preceding atrophy of the hippocampus. In the current study, we have performed a meta-analysis of publicly available expression data sets of the entorhinal cortex (EC) in order to identify potential pathways underlying AD pathology. The meta-analysis identified 1915 differentially expressed genes (DEGs) between the EC from normal and AD patients. Among the downregulated DEGs, we found a significant enrichment of biological processes pertaining to the “neuronal system” (R-HSA-112316) and the “synaptic signaling” (GO:0099536), while the “regulation of protein catabolic process” (GO:00042176) and “transport of small molecules” (R-HSA-382551) resulted in enrichment among both the upregulated and downregulated DEGs. Finally, by means of an in silico pharmacology approach, we have prioritized drugs and molecules potentially able to revert the transcriptional changes associated with AD pathology. The drugs with a mostly anti-correlated signature were: efavirenz, an anti-retroviral drug; tacrolimus, a calcineurin inhibitor; and sirolimus, an mTOR inhibitor. Among the predicted drugs, those potentially able to cross the blood-brain barrier have also been identified. Overall, our study found a disease-specific set of dysfunctional biological pathways characterizing the EC in AD patients and identified a set of drugs that could in the future be exploited as potential therapeutic strategies. The approach used in the current study has some limitations, as it does not account for possible post-transcriptional events regulating the cellular phenotype, and also, much clinical information about the samples included in the meta-analysis was not available. However, despite these limitations, our study sets the basis for future investigations on the pathogenetic processes occurring in AD and proposes the repurposing of currently used drugs for the treatment of AD patients.

DTI-HETA: prediction of drug-target interactions based on GCN and GAT on heterogeneous graph

Drug-target interaction (DTI) prediction plays an important role in drug repositioning, drug discovery and drug design. However, due to the large size of the chemical and genomic spaces and the complex interactions between drugs and targets, experimental identification of DTIs is costly and time-consuming. In recent years, the emerging graph neural network (GNN) has been applied to DTI prediction because DTIs can be represented effectively using graphs. However, some of these methods are only based on homogeneous graphs, and some consist of two decoupled steps that cannot be trained jointly. To further explore GNN-based DTI prediction by integrating heterogeneous graph information, this study regards DTI prediction as a link prediction problem and proposes an end-to-end model based on HETerogeneous graph with Attention mechanism (DTI-HETA). In this model, a heterogeneous graph is first constructed based on the drug-drug and target-target similarity matrices and the DTI matrix. Then, the graph convolutional neural network is utilized to obtain the embedded representation of the drugs and targets. To highlight the contribution of different neighborhood nodes to the central node in aggregating the graph convolution information, a graph attention mechanism is introduced into the node embedding process. Afterward, an inner product decoder is applied to predict DTIs. To evaluate the performance of DTI-HETA, experiments are conducted on two datasets. The experimental results show that our model is superior to the state-of-the-art methods. Also, the identification of novel DTIs indicates that DTI-HETA can serve as a powerful tool for integrating heterogeneous graph information to predict DTIs.

Drug target inference by mining transcriptional data using a novel graph convolutional network framework

A fundamental challenge that arises in biomedicine is the need to characterize compounds in a relevant cellular context in order to reveal potential on-target or off-target effects. Recently, the fast accumulation of gene transcriptional profiling data provides us an unprecedented opportunity to explore the protein targets of chemical compounds from the perspective of cell transcriptomics and RNA biology. Here, we propose a novel Siamese spectral-based graph convolutional network (SSGCN) model for inferring the protein targets of chemical compounds from gene transcriptional profiles. Although the gene signature of a compound perturbation only provides indirect clues of the interacting targets, and the biological networks under different experiment conditions further complicate the situation, the SSGCN model was successfully trained to learn from known compound-target pairs by uncovering the hidden correlations between compound perturbation profiles and gene knockdown profiles. On a benchmark set and a large time-split validation dataset, the model achieved higher target inference accuracy as compared to previous methods such as Connectivity Map. Further experimental validations of prediction results highlight the practical usefulness of SSGCN in either inferring the interacting targets of compound, or reversely, in finding novel inhibitors of a given target of interest.

A Study on Methodologies of Drug Repositioning Using Biomedical Big Data: A Focus on Diabetes Mellitus

Drug repositioning is a strategy for identifying new applications of an existing drug that has been previously proven to be safe. Based on several examples of drug repositioning, we aimed to determine the methodologies and relevant steps associated with drug repositioning that should be pursued in the future. Reports on drug repositioning, retrieved from PubMed from January 2011 to December 2020, were classified based on an analysis of the methodology and reviewed by experts. Among various drug repositioning methods, the network-based approach was the most common (38.0%, 186/490 cases), followed by machine learning/deep learningbased (34.3%, 168/490 cases), text mining-based (7.1%, 35/490 cases), semantic-based (5.3%, 26/490 cases), and others (15.3%, 75/490 cases). Although drug repositioning offers several advantages, its implementation is curtailed by the need for prior, conclusive clinical proof. This approach requires the construction of various databases, and a deep understanding of the process underlying repositioning is quintessential. An in-depth understanding of drug repositioning could reduce the time, cost, and risks inherent to early drug development, providing reliable scientific evidence. Furthermore, regarding patient safety, drug repurposing might allow the discovery of new relationships between drugs and diseases.

Multiview network embedding for drug-target Interactions prediction by consistent and complementary information preserving

Accurate prediction of drug-target interactions (DTIs) can reduce the cost and time of drug repositioning and drug discovery. Many current methods integrate information from multiple data sources of drug and target to improve DTIs prediction accuracy. However, these methods do not consider the complex relationship between different data sources. In this study, we propose a novel computational framework, called MccDTI, to predict the potential DTIs by multiview network embedding, which can integrate the heterogenous information of drug and target. MccDTI learns high-quality low-dimensional representations of drug and target by preserving the consistent and complementary information between multiview networks. Then MccDTI adopts matrix completion scheme for DTIs prediction based on drug and target representations. Experimental results on two datasets show that the prediction accuracy of MccDTI outperforms four state-of-the-art methods for DTIs prediction. Moreover, literature verification for DTIs prediction shows that MccDTI can predict the reliable potential DTIs. These results indicate that MccDTI can provide a powerful tool to predict new DTIs and accelerate drug discovery. The code and data are available at: https://github.com/ShangCS/MccDTI.

Combining CRISPRi and metabolomics for functional annotation of compound libraries

Molecular profiling of small-molecules offers invaluable insights into the function of compounds and allows for hypothesis generation about small molecule direct targets and secondary effects. However, current profiling methods are either limited in the number of measurable parameters or throughput. Here, we developed a multiplexed, unbiased framework that, by linking genetic to drug-induced changes in nearly a thousand metabolites, allows for high-throughput functional annotation of compound libraries in Escherichia coli. First, we generated a reference map of metabolic changes from (CRISPR) interference with 352 genes in all major essential biological processes. Next, based on the comparison of genetic with 1342 drug-induced metabolic changes we made de novo predictions of compound functionality and revealed antibacterials with unconventional Modes of Action. We show that our framework, combining dynamic gene silencing with metabolomics, can be adapted as a general strategy for comprehensive high-throughput analysis of compound functionality, from bacteria to human cell lines.

A survey of optimal strategy for signature-based drug repositioning and an application to liver cancer

Pharmacologic perturbation projects, such as Connectivity Map (CMap) and Library of Integrated Network-based Cellular Signatures (LINCS), have produced many perturbed expression data, providing enormous opportunities for computational therapeutic discovery. However, there is no consensus on which methodologies and parameters are the most optimal to conduct such analysis. Aiming to fill this gap, new benchmarking standards were developed to quantitatively evaluate drug retrieval performance. Investigations of potential factors influencing drug retrieval were conducted based on these standards. As a result, we determined an optimal approach for LINCS data-based therapeutic discovery. With this approach, homoharringtonine (HHT) was identified to be a candidate agent with potential therapeutic and preventive effects on liver cancer. The antitumor and antifibrotic activity of HHT was validated experimentally using subcutaneous xenograft tumor model and carbon tetrachloride (CCL₄)-induced liver fibrosis model, demonstrating the reliability of the prediction results. In summary, our findings will not only impact the future applications of LINCS data but also offer new opportunities for therapeutic intervention of liver cancer.

Identification of Nifurtimox and Chrysin as Anti-Influenza Virus Agents by Clinical Transcriptome Signature Reversion

The rapid development in the field of transcriptomics provides remarkable biomedical insights for drug discovery. In this study, a transcriptome signature reversal approach was conducted to identify the agents against influenza A virus (IAV) infection through dissecting gene expression changes in response to disease or compounds’ perturbations. Two compounds, nifurtimox and chrysin, were identified by a modified Kolmogorov–Smirnov test statistic based on the transcriptional signatures from 81 IAV-infected patients and the gene expression profiles of 1309 compounds. Their activities were verified in vitro with half maximal effective concentrations (EC₅₀s) from 9.1 to 19.1 μM against H1N1 or H3N2. It also suggested that the two compounds interfered with multiple sessions in IAV infection by reversing the expression of 28 IAV informative genes. Through network-based analysis of the 28 reversed IAV informative genes, a strong synergistic effect of the two compounds was revealed, which was confirmed in vitro. By using the transcriptome signature reversion (TSR) on clinical datasets, this study provides an efficient scheme for the discovery of drugs targeting multiple host factors regarding clinical signs and symptoms, which may also confer an opportunity for decelerating drug-resistant variant emergence.

ALDPI: adaptively learning importance of multi-scale topologies and multi-modality similarities for drug-protein interaction prediction

MOTIVATION

Effective computational methods to predict drug-protein interactions (DPIs) are vital for drug discovery in reducing the time and cost of drug development. Recent DPI prediction methods mainly exploit graph data composed of multiple kinds of connections among drugs and proteins. Each node in the graph usually has topological structures with multiple scales formed by its first-order neighbors and multi-order neighbors. However, most of the previous methods do not consider the topological structures of multi-order neighbors. In addition, deep integration of the multi-modality similarities of drugs and proteins is also a challenging task.

RESULTS

We propose a model called ALDPI to adaptively learn the multi-scale topologies and multi-modality similarities with various significance levels. We first construct a drug-protein heterogeneous graph, which is composed of the interactions and the similarities with multiple modalities among drugs and proteins. An adaptive graph learning module is then designed to learn important kinds of connections in heterogeneous graph and generate new topology graphs. A module based on graph convolutional autoencoders is established to learn multiple representations, which imply the node attributes and multiple-scale topologies composed of one-order and multi-order neighbors, respectively. We also design an attention mechanism at neighbor topology level to distinguish the importance of these representations. Finally, since each similarity modality has its specific features, we construct a multi-layer convolutional neural network-based module to learn and fuse multi-modality features to obtain the attribute representation of each drug-protein node pair. Comprehensive experimental results show ALDPI's superior performance over six state-of-the-art methods. The results of recall rates of top-ranked candidates and case studies on five drugs further demonstrate the ability of ALDPI to discover potential drug-related protein candidates.

CONTACT

zhang@hlju.edu.cn.

Causal reasoning over knowledge graphs leveraging drug-perturbed and disease-specific transcriptomic signatures for drug discovery

Network-based approaches are becoming increasingly popular for drug discovery as they provide a systems-level overview of the mechanisms underlying disease pathophysiology. They have demonstrated significant early promise over other methods of biological data representation, such as in target discovery, side effect prediction and drug repurposing. In parallel, an explosion of -omics data for the deep characterization of biological systems routinely uncovers molecular signatures of disease for similar applications. Here, we present RPath, a novel algorithm that prioritizes drugs for a given disease by reasoning over causal paths in a knowledge graph (KG), guided by both drug-perturbed as well as disease-specific transcriptomic signatures. First, our approach identifies the causal paths that connect a drug to a particular disease. Next, it reasons over these paths to identify those that correlate with the transcriptional signatures observed in a drug-perturbation experiment, and anti-correlate to signatures observed in the disease of interest. The paths which match this signature profile are then proposed to represent the mechanism of action of the drug. We demonstrate how RPath consistently prioritizes clinically investigated drug-disease pairs on multiple datasets and KGs, achieving better performance over other similar methodologies. Furthermore, we present two case studies showing how one can deconvolute the predictions made by RPath as well as predict novel targets.

A community challenge for a pancancer drug mechanism of action inference from perturbational profile data

The Columbia Cancer Target Discovery and Development (CTD2) Center is developing PANACEA, a resource comprising dose-responses and RNA sequencing (RNA-seq) profiles of 25 cell lines perturbed with ∼400 clinical oncology drugs, to study a tumor-specific drug mechanism of action. Here, this resource serves as the basis for a DREAM Challenge assessing the accuracy and sensitivity of computational algorithms for de novo drug polypharmacology predictions. Dose-response and perturbational profiles for 32 kinase inhibitors are provided to 21 teams who are blind to the identity of the compounds. The teams are asked to predict high-affinity binding targets of each compound among ∼1,300 targets cataloged in DrugBank. The best performing methods leverage gene expression profile similarity analysis as well as deep-learning methodologies trained on individual datasets. This study lays the foundation for future integrative analyses of pharmacogenomic data, reconciliation of polypharmacology effects in different tumor contexts, and insights into network-based assessments of drug mechanisms of action.

•

Drug-perturbed RNA sequencing data can be used to identify drug targets

•

Technology-based drug-target definitions often subsume literature definitions

•

Literature and screening datasets provide complementary information on drug mechanisms

High ORAI3 expression correlates with good prognosis in human muscle-invasive bladder cancer

The involvement of store-operated calcium channels (SOCCs) in tumor initiation and metastatic dissemination has been extensively studied, but how its member ORAI3 influences tumor progression is still elusive. The present study aimed to evaluate the prognostic value of ORAI3 expression and examine the correlation between ORAI3 expression and immune cell infiltration within the tumor microenvironment (TME) in human muscle-invasive bladder cancer (MIBC). We examined the expression profile of ORAI3 in MIBC using data from two databases; analyzed the correlation between ORAI3 expression and patient survival; explored cellular pathways related to ORAI3 expression by Gene Set Enrichment Analysis (GSEA); and predicted potential drugs using Connectivity Map (CMap). ORAI3 was significantly lower expressed in tumor mass compared to normal samples in MIBC, with a higher level of methylation at the promoter region in tumor than in normal tissue, indicating that ORAI3 is suppressed during cancer progression. Survival analysis showed that higher expression of ORAI3 correlated with good prognosis in MIBC. GSEA demonstrated that ORAI3 expression inversely correlated with cell differentiation, development and gene silencing, with differential expression of genes involved in epidermal and keratinocyte differentiation pathways and inflammatory responses. RNA sequencing of an ORAI3-silenced human bladder cancer cell line (T24 cells) corroborated enhancement of pro-neoplastic pathways in absence of ORAI3. Western blottingMoreover, ORAI3 facilitated the recruitment of Th17 cells and natural killer cells, whereas hampered Th2 and macrophage infiltration. Our results revealed 4 molecules with potential to be beneficial as adjuvant drugs in MIBC treatment. We concluded that high ORAI3 expression correlates with increased survival in human MIBC.

Transcription Factor Activation Profiles (TFAP) identify compounds promoting differentiation of Acute Myeloid Leukemia cell lines

Repurposing of drugs for new therapeutic use has received considerable attention for its potential to limit time and cost of drug development. Here we present a new strategy to identify chemicals that are likely to promote a desired phenotype. We used data from the Connectivity Map (CMap) to produce a ranked list of drugs according to their potential to activate transcription factors that mediate myeloid differentiation of leukemic progenitor cells. To validate our strategy, we tested the in vitro differentiation potential of candidate compounds using the HL-60 human cell line as a myeloid differentiation model. Ten out of 22 compounds, which were ranked high in the inferred list, were confirmed to promote significant differentiation of HL-60. These compounds may be considered candidate for differentiation therapy. The method that we have developed is versatile and it can be adapted to different drug repurposing projects.

Repurposing – second life for drugs

Drug repurposing refers to finding new indications for existing drugs. The paradigm shift from traditional drug discovery to drug repurposing is driven by the fact that new drug pipelines are getting dried up because of mounting Research & Development (R&D) costs, long timeline for new drug development, low success rate for new molecular entities, regulatory hurdles coupled with revenue loss from patent expiry and competition from generics. Anaemic drug pipelines along with increasing demand for newer effective, cheaper, safer drugs and unmet medical needs call for new strategies of drug discovery and, drug repurposing seems to be a promising avenue for such endeavours. Drug repurposing strategies have progressed over years from simple serendipitous observations to more complex computational methods in parallel with our ever-growing knowledge on drugs, diseases, protein targets and signalling pathways but still the knowledge is far from complete. Repurposed drugs too have to face many obstacles, although lesser than new drugs, before being successful.

Graph Convolutional Autoencoder and Generative Adversarial Network-Based Method for Predicting Drug-Target Interactions

The computational prediction of novel drug-target interactions (DTIs) may effectively speed up the process of drug repositioning and reduce its costs. Most previous methods integrated multiple kinds of connections about drugs and targets by constructing shallow prediction models. These methods failed to deeply learn the low-dimension feature vectors for drugs and targets and ignored the distribution of these feature vectors. We proposed a graph convolutional autoencoder and generative adversarial network (GAN)-based method, GANDTI, to predict DTIs. We constructed a drug-target heterogeneous network to integrate various connections related to drugs and targets, i.e., the similarities and interactions between drugs or between targets and the interactions between drugs and targets. A graph convolutional autoencoder was established to learn the network embeddings of the drug and target nodes in a low-dimensional feature space, and the autoencoder deeply integrated different kinds of connections within the network. A GAN was introduced to regularize the feature vectors of nodes into a Gaussian distribution. Severe class imbalance exists between known and unknown DTIs. Thus, we constructed a classifier based on an ensemble learning model, LightGBM, to estimate the interaction propensities of drugs and targets. This classifier completely exploited all unknown DTIs and counteracted the negative effect of class imbalance. The experimental results indicated that GANDTI outperforms several state-of-the-art methods for DTI prediction. Additionally, case studies of five drugs demonstrated the ability of GANDTI to discover the potential targets for drugs.

AI-powered drug repurposing for developing COVID-19 treatments

Emerging infectious diseases are an ever-present threat to public health, and COVID-19 is the most recent example. There is an urgent need to develop a robust framework to combat the disease with safe and effective therapeutic options. Compared to de novo drug discovery, drug repurposing may offer a lower-cost and faster drug discovery paradigm to explore potential treatment options of existing drugs. This chapter elucidates the advantages of artificial intelligence (AI) in enhancing the drug repurposing process from a data science perspective, using COVID-19 as an example. First, we elaborate on how AI-powered drug repurposing benefits from the accumulated data and knowledge of COVID-19 natural history and pathogenesis. Second, we summarize the pros and cons of AI-powered drug repurposing strategies to facilitate fit-for-purpose selection. Finally, we outline challenges of AI-powered drug repurposing from a regulatory perspective and suggest some potential solutions.

•

AI-powered drug purposing is promising for emerging treatments for COVID-19 infection.

•

Accumulated biological data profiles facilitate AI-based drug repurposing efforts for development of COVID-19 therapies.

•

The ‘fit-for-purpose selection of AI-powered drug repurposing strategies is key to uncovering hidden information among drugs, targets, and diseases.

•

Efforts from different stakeholders boost the adoption of AI-powered drug repurposing in the regulatory setting.

Beyondcell: targeting cancer therapeutic heterogeneity in single-cell RNA-seq data

We present Beyondcell, a computational methodology for identifying tumour cell subpopulations with distinct drug responses in single-cell RNA-seq data and proposing cancer-specific treatments. Our method calculates an enrichment score in a collection of drug signatures, delineating therapeutic clusters (TCs) within cellular populations. Additionally, Beyondcell determines the therapeutic differences among cell populations and generates a prioritised sensitivity-based ranking in order to guide drug selection. We performed Beyondcell analysis in five single-cell datasets and demonstrated that TCs can be exploited to target malignant cells both in cancer cell lines and tumour patients. Beyondcell is available at: https://gitlab.com/bu_cnio/beyondcell .

Exposure to GenX and its novel analogs disrupts fatty acid metabolism in male mice

Perfluoroalkyl ether carboxylic acids (PFECAs), including hexafluoropropylene oxide dimer acid (HFPO-DA, GenX), have been widely used as alternatives to perfluorooctanoic acid (PFOA) and subsequently detected in various environmental matrices. Despite this, public information regarding their hepatotoxicity remains limited. Here, to compare the hepatotoxicity of PFECAs and identify better alternatives for GenX, adult male mice were exposed to different concentrations (0.4, 2, and 10 mg/kg/d) of PFOA, GenX, and its analogs (PFMO2HpA and PFMO3NA) for 28 d. Results demonstrated increased hepatomegaly and disturbed fatty acid metabolism with increasing treatment doses. After dimensionality reduction analysis, significant differences were observed in the relative liver weights and liver and serum biochemical parameters among the four clusters. Furthermore, when chemical concentrations in the liver were similar, no differences in the indicators of liver injury associated with fatty acid metabolism were observed among groups in the same clusters. Our results suggest that dimensionality reduction analysis is a useful strategy for analyzing samples exposed to multiple compounds at different doses. Furthermore, PFECAs exhibit similar hepatotoxicities at the same cumulative hepatic concentration in mice with constant body weight, while PFMO2HpA exhibits lower hepatotoxicity compared to GenX at the same dose.

Drug Repurposing Using Modularity Clustering in Drug-Drug Similarity Networks Based on Drug-Gene Interactions

Drug repurposing is a valuable alternative to traditional drug design based on the assumption that medicines have multiple functions. Computer-based techniques use ever-growing drug databases to uncover new drug repurposing hints, which require further validation with in vitro and in vivo experiments. Indeed, such a scientific undertaking can be particularly effective in the case of rare diseases (resources for developing new drugs are scarce) and new diseases such as COVID-19 (designing new drugs require too much time). This paper introduces a new, completely automated computational drug repurposing pipeline based on drug-gene interaction data. We obtained drug-gene interaction data from an earlier version of DrugBank, built a drug-gene interaction network, and projected it as a drug-drug similarity network (DDSN). We then clustered DDSN by optimizing modularity resolution, used the ATC codes distribution within each cluster to identify potential drug repurposing candidates, and verified repurposing hints with the latest DrugBank ATC codes. Finally, using the best modularity resolution found with our method, we applied our pipeline to the latest DrugBank drug-gene interaction data to generate a comprehensive drug repurposing hint list.

Predicting drug-drug interactions by graph convolutional network with multi-kernel

Drug repositioning is proposed to find novel usages for existing drugs. Among many types of drug repositioning approaches, predicting drug-drug interactions (DDIs) helps explore the pharmacological functions of drugs and achieves potential drugs for novel treatments. A number of models have been applied to predict DDIs. The DDI network, which is constructed from the known DDIs, is a common part in many of the existing methods. However, the functions of DDIs are different, and thus integrating them in a single DDI graph may overlook some useful information. We propose a graph convolutional network with multi-kernel (GCNMK) to predict potential DDIs. GCNMK adopts two DDI graph kernels for the graph convolutional layers, namely, increased DDI graph consisting of 'increase'-related DDIs and decreased DDI graph consisting of 'decrease'-related DDIs. The learned drug features are fed into a block with three fully connected layers for the DDI prediction. We compare various types of drug features, whereas the target feature of drugs outperforms all other types of features and their concatenated features. In comparison with three different DDI prediction methods, our proposed GCNMK achieves the best performance in terms of area under receiver operating characteristic curve and area under precision-recall curve. In case studies, we identify the top 20 potential DDIs from all unknown DDIs, and the top 10 potential DDIs from the unknown DDIs among breast, colorectal and lung neoplasms-related drugs. Most of them have evidence to support the existence of their interactions. fangxiang.wu@usask.ca.

Drug Repurposing in Rare Diseases: An Integrative Study of Drug Screening and Transcriptomic Analysis in Nephropathic Cystinosis

Diagnosis and cure for rare diseases represent a great challenge for the scientific community who often comes up against the complexity and heterogeneity of clinical picture associated to a high cost and time-consuming drug development processes. Here we show a drug repurposing strategy applied to nephropathic cystinosis, a rare inherited disorder belonging to the lysosomal storage diseases. This approach consists in combining mechanism-based and cell-based screenings, coupled with an affordable computational analysis, which could result very useful to predict therapeutic responses at both molecular and system levels. Then, we identified potential drugs and metabolic pathways relevant for the pathophysiology of nephropathic cystinosis by comparing gene-expression signature of drugs that share common mechanisms of action or that involve similar pathways with the disease gene-expression signature achieved with RNA-seq.

GVDTI: graph convolutional and variational autoencoders with attribute-level attention for drug-protein interaction prediction

MOTIVATION

Identifying proteins that interact with drugs plays an important role in the initial period of developing drugs, which helps to reduce the development cost and time. Recent methods for predicting drug-protein interactions mainly focus on exploiting various data about drugs and proteins. These methods failed to completely learn and integrate the attribute information of a pair of drug and protein nodes and their attribute distribution.

RESULTS

We present a new prediction method, GVDTI, to encode multiple pairwise representations, including attention-enhanced topological representation, attribute representation and attribute distribution. First, a framework based on graph convolutional autoencoder is constructed to learn attention-enhanced topological embedding that integrates the topology structure of a drug-protein network for each drug and protein nodes. The topological embeddings of each drug and each protein are then combined and fused by multi-layer convolution neural networks to obtain the pairwise topological representation, which reveals the hidden topological relationships between drug and protein nodes. The proposed attribute-wise attention mechanism learns and adjusts the importance of individual attribute in each topological embedding of drug and protein nodes. Secondly, a tri-layer heterogeneous network composed of drug, protein and disease nodes is created to associate the similarities, interactions and associations across the heterogeneous nodes. The attribute distribution of the drug-protein node pair is encoded by a variational autoencoder. The pairwise attribute representation is learned via a multi-layer convolutional neural network to deeply integrate the attributes of drug and protein nodes. Finally, the three pairwise representations are fused by convolutional and fully connected neural networks for drug-protein interaction prediction. The experimental results show that GVDTI outperformed other seven state-of-the-art methods in comparison. The improved recall rates indicate that GVDTI retrieved more actual drug-protein interactions in the top ranked candidates than conventional methods. Case studies on five drugs further confirm GVDTI's ability in discovering the potential candidate drug-related proteins.

CONTACT

zhang@hlju.edu.cn  Supplementary information: Supplementary data are available at Briefings in Bioinformatics online.

Learning multi-scale heterogeneous representations and global topology for drug-target interaction prediction

Identification of drug-target interactions (DTIs) plays a critical role in drug discovery and repositioning. Deep integration of inter-connections and intra-similarities between heterogeneous multi-source data related to drugs and targets, however, is a challenging issue. We propose a DTI prediction model by learning from drug and protein related multi-scale attributes and global topology formed by heterogeneous connections. A drug-protein-disease heterogeneous network (RPD-Net) is firstly constructed to associate diverse similarities, interactions and associations across nodes. Secondly, we propose a multi-scale pairwise deep representation learning module consisting of a new embedding strategy to integrate diverse inter-relations and intra-relations, and dilation convolutions for multi-scale deep representation extraction. A global topology learning module is proposed which is composed of strategy based on non-negative matrix factorization (NMF) to extract topology from RPD-Net, and a new relational-level attention mechanism for discriminative topology embedding. Experimental results using public dataset demonstrate improved performance over state-of-the-art methods and contributions of our major innovations. Evaluation results by top k recall rates and case studies on five drugs further show the effectiveness in retrieving potential target candidates for drugs.

Human Protein Complex-Based Drug Signatures for Personalized Cancer Medicine

Disease signature-based drug repositioning approaches typically first identify a disease signature from gene expression profiles of disease samples to represent a particular disease. Then such a disease signature is connected with the drug-induced gene expression profiles to find potential drugs for the particular disease. In order to obtain reliable disease signatures, the size of disease samples should be large enough, which is not always a single case in practice, especially for personalized medicine. On the other hand, the sample sizes of drug-induced gene expression profiles are generally large. In this study, we propose a new drug repositioning approach (HDgS), in which the drug signature is first identified from drug-induced gene expression profiles, and then connected to the gene expression profiles of disease samples to find the potential drugs for patients. In order to take the dependencies among genes into account, the human protein complexes (HPC) are used to define the drug signature. The proposed HDgS is applied to the drug-induced gene expression profiles in LINCS and several types of cancer samples. The results indicate that the HPC-based drug signature can effectively find drug candidates for patients and that the proposed HDgS can be applied for personalized medicine with even one patient sample.

DeepCellState: An autoencoder-based framework for predicting cell type specific transcriptional states induced by drug treatment

Drug treatment induces cell type specific transcriptional programs, and as the number of combinations of drugs and cell types grows, the cost for exhaustive screens measuring the transcriptional drug response becomes intractable. We developed DeepCellState, a deep learning autoencoder-based framework, for predicting the induced transcriptional state in a cell type after drug treatment, based on the drug response in another cell type. Training the method on a large collection of transcriptional drug perturbation profiles, prediction accuracy improves significantly over baseline and alternative deep learning approaches when applying the method to two cell types, with improved accuracy when generalizing the framework to additional cell types. Treatments with drugs or whole drug families not seen during training are predicted with similar accuracy, and the same framework can be used for predicting the results from other interventions, such as gene knock-downs. Finally, analysis of the trained model shows that the internal representation is able to learn regulatory relationships between genes in a fully data-driven manner.

TRIOMPHE: Transcriptome-Based Inference and Generation of Molecules with Desired Phenotypes by Machine Learning

One of the most challenging tasks in the drug-discovery process is the efficient identification of small molecules with desired phenotypes. In this study, we propose a novel computational method for omics-based de novo drug design, which we call TRIOMPHE (transcriptome-based inference and generation of molecules with desired phenotypes). We investigated the correlation between chemically induced transcriptome profiles (reflecting cellular responses to compound treatment) and genetically perturbed transcriptome profiles (reflecting cellular responses to gene knock-down or gene overexpression of target proteins) in terms of ligand-target interactions. Subsequently, we developed novel machine learning methods to generate the chemical structures of new molecules with desired transcriptome profiles in the framework of a variational autoencoder. The use of desired transcriptome profiles enables the automatic design of molecules that are likely to have bioactivities for target proteins of interest. We showed that our methods can generate chemically valid molecules that are likely to have biological activities on 10 target proteins; moreover, they can outperform previous methods that had the same objective. Our omics-based structure generator is expected to be useful for the de novo design of drugs for a variety of target proteins.

Validation strategies for target prediction methods

Computational methods for target prediction, based on molecular similarity and network-based approaches, machine learning, docking and others, have evolved as valuable and powerful tools to aid the challenging task of mode of action identification for bioactive small molecules such as drugs and drug-like compounds. Critical to discerning the scope and limitations of a target prediction method is understanding how its performance was evaluated and reported. Ideally, large-scale prospective experiments are conducted to validate the performance of a model; however, this expensive and time-consuming endeavor is often not feasible. Therefore, to estimate the predictive power of a method, statistical validation based on retrospective knowledge is commonly used. There are multiple statistical validation techniques that vary in rigor. In this review we discuss the validation strategies employed, highlighting the usefulness and constraints of the validation schemes and metrics that are employed to measure and describe performance. We address the limitations of measuring only generalized performance, given that the underlying bioactivity and structural data are biased towards certain small-molecule scaffolds and target families, and suggest additional aspects of performance to consider in order to produce more detailed and realistic estimates of predictive power. Finally, we describe the validation strategies that were employed by some of the most thoroughly validated and accessible target prediction methods.

Predicting Drug-Induced Liver Injury Using Machine Learning on a Diverse Set of Predictors

A major challenge in drug development is safety and toxicity concerns due to drug side effects. One such side effect, drug-induced liver injury (DILI), is considered a primary factor in regulatory clearance. The Critical Assessment of Massive Data Analysis (CAMDA) 2020 CMap Drug Safety Challenge goal was to develop prediction models based on gene perturbation of six preselected cell-lines (CMap L1000), extended structural information (MOLD2), toxicity data (TOX21), and FDA reporting of adverse events (FAERS). Four types of DILI classes were targeted, including two clinically relevant scores and two control classifications, designed by the CAMDA organizers. The L1000 gene expression data had variable drug coverage across cell lines with only 247 out of 617 drugs in the study measured in all six cell types. We addressed this coverage issue by using Kru-Bor ranked merging to generate a singular drug expression signature across all six cell lines. These merged signatures were then narrowed down to the top and bottom 100, 250, 500, or 1,000 genes most perturbed by drug treatment. These signatures were subject to feature selection using Fisher's exact test to identify genes predictive of DILI status. Models based solely on expression signatures had varying results for clinical DILI subtypes with an accuracy ranging from 0.49 to 0.67 and Matthews Correlation Coefficient (MCC) values ranging from -0.03 to 0.1. Models built using FAERS, MOLD2, and TOX21 also had similar results in predicting clinical DILI scores with accuracy ranging from 0.56 to 0.67 with MCC scores ranging from 0.12 to 0.36. To incorporate these various data types with expression-based models, we utilized soft, hard, and weighted ensemble voting methods using the top three performing models for each DILI classification. These voting models achieved a balanced accuracy up to 0.54 and 0.60 for the clinically relevant DILI subtypes. Overall, from our experiment, traditional machine learning approaches may not be optimal as a classification method for the current data.

Drug-target interaction predication via multi-channel graph neural networks

Drug-target interaction (DTI) is an important step in drug discovery. Although there are many methods for predicting drug targets, these methods have limitations in using discrete or manual feature representations. In recent years, deep learning methods have been used to predict DTIs to improve these defects. However, most of the existing deep learning methods lack the fusion of topological structure and semantic information in DPP representation learning process. Besides, when learning the DPP node representation in the DPP network, the different influences between neighboring nodes are ignored. In this paper, a new model DTI-MGNN based on multi-channel graph convolutional network and graph attention is proposed for DTI prediction. We use two independent graph attention networks to learn the different interactions between nodes for the topology graph and feature graph with different strengths. At the same time, we use a graph convolutional network with shared weight matrices to learn the common information of the two graphs. The DTI-MGNN model combines topological structure and semantic features to improve the representation learning ability of DPPs, and obtain the state-of-the-art results on public datasets. Specifically, DTI-MGNN has achieved a high accuracy in identifying DTIs (the area under the receiver operating characteristic curve is 0.9665).

Identification of new target proteins of a Urotensin-II receptor antagonist using transcriptome-based drug repositioning approach

Drug repositioning research using transcriptome data has recently attracted attention. In this study, we attempted to identify new target proteins of the urotensin-II receptor antagonist, KR-37524 (4-(3-bromo-4-(piperidin-4-yloxy)benzyl)-N-(3-(dimethylamino)phenyl)piperazine-1-carboxamide dihydrochloride), using a transcriptome-based drug repositioning approach. To do this, we obtained KR-37524-induced gene expression profile changes in four cell lines (A375, A549, MCF7, and PC3), and compared them with the approved drug-induced gene expression profile changes available in the LINCS L1000 database to identify approved drugs with similar gene expression profile changes. Here, the similarity between the two gene expression profile changes was calculated using the connectivity score. We then selected proteins that are known targets of the top three approved drugs with the highest connectivity score in each cell line (12 drugs in total) as potential targets of KR-37524. Seven potential target proteins were experimentally confirmed using an in vitro binding assay. Through this analysis, we identified that neurologically regulated serotonin transporter proteins are new target proteins of KR-37524. These results indicate that the transcriptome-based drug repositioning approach can be used to identify new target proteins of a given compound, and we provide a standalone software developed in this study that will serve as a useful tool for drug repositioning.

Drug Discovery of Spinal Muscular Atrophy (SMA) from the Computational Perspective: A Comprehensive Review

Spinal muscular atrophy (SMA), one of the leading inherited causes of child mortality, is a rare neuromuscular disease arising from loss-of-function mutations of the survival motor neuron 1 (SMN1) gene, which encodes the SMN protein. When lacking the SMN protein in neurons, patients suffer from muscle weakness and atrophy, and in the severe cases, respiratory failure and death. Several therapeutic approaches show promise with human testing and three medications have been approved by the U.S. Food and Drug Administration (FDA) to date. Despite the shown promise of these approved therapies, there are some crucial limitations, one of the most important being the cost. The FDA-approved drugs are high-priced and are shortlisted among the most expensive treatments in the world. The price is still far beyond affordable and may serve as a burden for patients. The blooming of the biomedical data and advancement of computational approaches have opened new possibilities for SMA therapeutic development. This article highlights the present status of computationally aided approaches, including in silico drug repurposing, network driven drug discovery as well as artificial intelligence (AI)-assisted drug discovery, and discusses the future prospects.

Topological network measures for drug repositioning

Drug repositioning has received increased attention since the past decade as several blockbuster drugs have come out of repositioning. Computational approaches are significantly contributing to these efforts, of which, network-based methods play a key role. Various structural (topological) network measures have thereby contributed to uncovering unintuitive functional relationships and repositioning candidates in drug-disease and other networks. This review gives a broad overview of the topic, and offers perspectives on the application of topological measures for network analysis. It also discusses unexplored measures, and draws attention to a wider scope of application efforts, especially in drug repositioning.

Transcriptional drug repositioning and cheminformatics approach for differentiation therapy of leukaemia cells

Differentiation therapy is attracting increasing interest in cancer as it can be more specific than conventional chemotherapy approaches, and it has offered new treatment options for some cancer types, such as treating acute promyelocytic leukaemia (APL) by retinoic acid. However, there is a pressing need to identify additional molecules which act in this way, both in leukaemia and other cancer types. In this work, we hence developed a novel transcriptional drug repositioning approach, based on both bioinformatics and cheminformatics components, that enables selecting such compounds in a more informed manner. We have validated the approach for leukaemia cells, and retrospectively retinoic acid was successfully identified using our method. Prospectively, the anti-parasitic compound fenbendazole was tested in leukaemia cells, and we were able to show that it can induce the differentiation of leukaemia cells to granulocytes in low concentrations of 0.1 μM and within as short a time period as 3 days. This work hence provides a systematic and validated approach for identifying small molecules for differentiation therapy in cancer.