Reconciling multiple connectivity scores for drug repurposing

A review on drug repurposing applicable to obesity

Obesity is a major public health concern and burden on individuals and healthcare systems. Due to the challenges and limitations of lifestyle adjustments, it is advisable to consider pharmacological treatment for people affected by obesity. However, the side effects and limited efficacy of available drugs make the obesity drug market far from sufficient. Drug repurposing involves identifying new applications for existing drugs and offers some advantages over traditional drug development approaches including lower costs and shorter development timelines. This review aims to provide an overview of drug repurposing for anti-obesity medications, including the rationale for repurposing, the challenges and approaches, and the potential drugs that are being investigated for repurposing. Through advanced computational techniques, researchers can unlock the potential of repurposed drugs to tackle the global obesity epidemic. Further research, clinical trials, and collaborative efforts are essential to fully explore and leverage the potential of drug repurposing in the fight against obesity.

Signature Search Polestar: a comprehensive drug repurposing method evaluation assistant for customized oncogenic signature

SUMMARY

The burgeoning high-throughput technologies have led to a significant surge in the scale of pharmacotranscriptomic datasets, especially for oncology. Signature search methods (SSMs), utilizing oncogenic signatures formed by differentially expressed genes through sequencing, have been instrumental in anti-cancer drug screening and identifying mechanisms of action without relying on prior knowledge. However, various studies have found that different SSMs exhibit varying performance across pharmacotranscriptomic datasets. In addition, the size of the oncogenic signature can also significantly impact the result of drug repurposing. Therefore, finding the optimal SSMs and customized oncogenic signature for a specific disease remains a challenge. To address this, we introduce Signature Search Polestar (SSP), a webserver integrating the largest pharmacotranscriptomic datasets of anti-cancer drugs from LINCS L1000 with five state-of-the-art SSMs (XSum, CMap, GSEA, ZhangScore, XCos). SSP provides three main modules: Benchmark, Robustness, and Application. Benchmark uses two indices, Area Under the Curve and Enrichment Score, based on drug annotations to evaluate SSMs at different oncogenic signature sizes. Robustness, applicable when drug annotations are insufficient, uses a performance score based on drug self-retrieval for evaluation. Application provides three screening strategies, single method, SS_all, and SS_cross, allowing users to freely utilize optimal SSMs with tailored oncogenic signature for drug repurposing.

AVAILABILITY AND IMPLEMENTATION

SSP is free at https://web.biotcm.net/SSP/. The current version of SSP is archived in https://doi.org/10.6084/m9.figshare.26524741.v1, allowing users to directly use or customize their own SSP webserver.

Integrated transcriptomics- and structure-based drug repositioning identifies drugs with proteasome inhibitor properties

Computational pharmacogenomics can potentially identify new indications for already approved drugs and pinpoint compounds with similar mechanism-of-action. Here, we used an integrated drug repositioning approach based on transcriptomics data and structure-based virtual screening to identify compounds with gene signatures similar to three known proteasome inhibitors (PIs; bortezomib, MG-132, and MLN-2238). In vitro validation of candidate compounds was then performed to assess proteasomal proteolytic activity, accumulation of ubiquitinated proteins, cell viability, and drug-induced expression in A375 melanoma and MCF7 breast cancer cells. Using this approach, we identified six compounds with PI properties ((-)-kinetin-riboside, manumycin-A, puromycin dihydrochloride, resistomycin, tegaserod maleate, and thapsigargin). Although the docking scores pinpointed their ability to bind to the β5 subunit, our in vitro study revealed that these compounds inhibited the β1, β2, and β5 catalytic sites to some extent. As shown with bortezomib, only manumycin-A, puromycin dihydrochloride, and tegaserod maleate resulted in excessive accumulation of ubiquitinated proteins and elevated HMOX1 expression. Taken together, our integrated drug repositioning approach and subsequent in vitro validation studies identified six compounds demonstrating properties similar to proteasome inhibitors.

Optimizing in silico drug discovery: simulation of connected differential expression signatures and applications to benchmarking

BACKGROUND

We present a novel simulation method for generating connected differential expression signatures. Traditional methods have struggled with the lack of reliable benchmarking data and biases in drug-disease pair labeling, limiting the rigorous benchmarking of connectivity-based approaches.

OBJECTIVE

Our aim is to develop a simulation method based on a statistical framework that allows for adjustable levels of parametrization, especially the connectivity, to generate a pair of interconnected differential signatures. This could help to address the issue of benchmarking data availability for connectivity-based drug repurposing approaches.

METHODS

We first detailed the simulation process and how it reflected real biological variability and the interconnectedness of gene expression signatures. Then, we generated several datasets to enable the evaluation of different existing algorithms that compare differential expression signatures, providing insights into their performance and limitations.

RESULTS

Our findings demonstrate the ability of our simulation to produce realistic data, as evidenced by correlation analyses and the log2 fold-change distribution of deregulated genes. Benchmarking reveals that methods like extreme cosine similarity and Pearson correlation outperform others in identifying connected signatures.

CONCLUSION

Overall, our method provides a reliable tool for simulating differential expression signatures. The data simulated by our tool encompass a wide spectrum of possibilities to challenge and evaluate existing methods to estimate connectivity scores. This may represent a critical gap in connectivity-based drug repurposing research because reliable benchmarking data are essential for assessing and advancing in the development of new algorithms. The simulation tool is available as a R package (General Public License (GPL) license) at https://github.com/cgonzalez-gomez/cosimu.

Combinatorial prediction of therapeutic perturbations using causally-inspired neural networks

As an alternative to target-driven drug discovery, phenotype-driven approaches identify compounds that counteract the overall disease effects by analyzing phenotypic signatures. Our study introduces a novel approach to this field, aiming to expand the search space for new therapeutic agents. We introduce PDGrapher, a causally-inspired graph neural network model designed to predict arbitrary perturbagens - sets of therapeutic targets - capable of reversing disease effects. Unlike existing methods that learn responses to perturbations, PDGrapher solves the inverse problem, which is to infer the perturbagens necessary to achieve a specific response - i.e., directly predicting perturbagens by learning which perturbations elicit a desired response. Experiments across eight datasets of genetic and chemical perturbations show that PDGrapher successfully predicted effective perturbagens in up to 9% additional test samples and ranked therapeutic targets up to 35% higher than competing methods. A key innovation of PDGrapher is its direct prediction capability, which contrasts with the indirect, computationally intensive models traditionally used in phenotypedriven drug discovery that only predict changes in phenotypes due to perturbations. The direct approach enables PDGrapher to train up to 30 times faster, representing a significant leap in efficiency. Our results suggest that PDGrapher can advance phenotype-driven drug discovery, offering a fast and comprehensive approach to identifying therapeutically useful perturbations.

PharmGWAS: a GWAS-based knowledgebase for drug repurposing

Leveraging genetics insights to promote drug repurposing has become a promising and active strategy in pharmacology. Indeed, among the 50 drugs approved by FDA in 2021, two-thirds have genetically supported evidence. In this regard, the increasing amount of widely available genome-wide association studies (GWAS) datasets have provided substantial opportunities for drug repurposing based on genetics discoveries. Here, we developed PharmGWAS, a comprehensive knowledgebase designed to identify candidate drugs through the integration of GWAS data. PharmGWAS focuses on novel connections between diseases and small-molecule compounds derived using a reverse relationship between the genetically-regulated expression signature and the drug-induced signature. Specifically, we collected and processed 1929 GWAS datasets across a diverse spectrum of diseases and 724 485 perturbation signatures pertaining to a substantial 33609 molecular compounds. To obtain reliable and robust predictions for the reverse connections, we implemented six distinct connectivity methods. In the current version, PharmGWAS deposits a total of 740 227 genetically-informed disease-drug pairs derived from drug-perturbation signatures, presenting a valuable and comprehensive catalog. Further equipped with its user-friendly web design, PharmGWAS is expected to greatly aid the discovery of novel drugs, the exploration of drug combination therapies and the identification of drug resistance or side effects. PharmGWAS is available at https://ngdc.cncb.ac.cn/pharmgwas.

DeepdlncUD: Predicting regulation types of small molecule inhibitors on modulating lncRNA expression by deep learning

Targeting lncRNAs by small molecules (SM-lncR) to alter their expression levels has emerged as an important therapeutic modality for disease treatment. To date, no computational tools have been dedicated to predicting small molecule-mediated upregulation or downregulation of lncRNA expression. Here, we introduce DeepdlncUD, which integrates predictions of nine deep learning algorithms together, to infer the regulation types of small molecules on modulating lncRNA expression. Through systematic optimization on a training set of 771 upregulation and 739 downregulation SM-lncR pairs, each encoding 1369 sequence, representational, and physiochemical features, this method outperforms a recently released program, DeepsmirUD, by achieving 0.674 in AUC (area under the receiver operating characteristic curve), 0.722 in AUCPR (area under the precision-recall curve), 0.681 in F1-score, and 0.516 in Jaccard Index on a test set of 222 SM-lncR pairs. By extracting 125 upregulation and 46 downregulation SM-lncR pairs that involve disease-associated lncRNAs, DeepdlncUD is shown to gain an accuracy of 0.700 in the pathological context. Using connectivity scores, around half of the small molecules are correctly estimated as drugs to treat lncRNA-regulated diseases. This tool can be run at a fast speed to assist the discovery of potential small molecule drugs of lncRNA targets on a large scale. DeepdlncUD is publicly available at https://github.com/2003100127/deepdlncud.

Small molecule-mediated targeting of microRNAs for drug discovery: Experiments, computational techniques, and disease implications

Small molecules have been providing medical breakthroughs for human diseases for more than a century. Recently, identifying small molecule inhibitors that target microRNAs (miRNAs) has gained importance, despite the challenges posed by labour-intensive screening experiments and the significant efforts required for medicinal chemistry optimization. Numerous experimentally-verified cases have demonstrated the potential of miRNA-targeted small molecule inhibitors for disease treatment. This new approach is grounded in their posttranscriptional regulation of the expression of disease-associated genes. Reversing dysregulated gene expression using this mechanism may help control dysfunctional pathways. Furthermore, the ongoing improvement of algorithms has allowed for the integration of computational strategies built on top of laboratory-based data, facilitating a more precise and rational design and discovery of lead compounds. To complement the use of extensive pharmacogenomics data in prioritising potential drugs, our previous work introduced a computational approach based on only molecular sequences. Moreover, various computational tools for predicting molecular interactions in biological networks using similarity-based inference techniques have been accumulated in established studies. However, there are a limited number of comprehensive reviews covering both computational and experimental drug discovery processes. In this review, we outline a cohesive overview of both biological and computational applications in miRNA-targeted drug discovery, along with their disease implications and clinical significance. Finally, utilizing drug-target interaction (DTIs) data from DrugBank, we showcase the effectiveness of deep learning for obtaining the physicochemical characterization of DTIs.

Cell-specific imputation of drug connectivity mapping with incomplete data

Drug repositioning allows expedited discovery of new applications for existing compounds, but re-screening vast compound libraries is often prohibitively expensive. "Connectivity mapping" is a process that links drugs to diseases by identifying compounds whose impact on expression in a collection of cells reverses the disease's impact on expression in disease-relevant tissues. The LINCS project has expanded the universe of compounds and cells for which data are available, but even with this effort, many clinically useful combinations are missing. To evaluate the possibility of repurposing drugs despite missing data, we compared collaborative filtering using either neighborhood-based or SVD imputation methods to two naive approaches via cross-validation. Methods were evaluated for their ability to predict drug connectivity despite missing data. Predictions improved when cell type was taken into account. Neighborhood collaborative filtering was the most successful method, with the best improvements in non-immortalized primary cells. We also explored which classes of compounds are most and least reliant on cell type for accurate imputation. We conclude that even for cells in which drug responses have not been fully characterized, it is possible to identify unassayed drugs that reverse in those cells the expression signatures observed in disease.

DeepsmirUD: Prediction of Regulatory Effects on microRNA Expression Mediated by Small Molecules Using Deep Learning

Aberrant miRNA expression has been associated with a large number of human diseases. Therefore, targeting miRNAs to regulate their expression levels has become an important therapy against diseases that stem from the dysfunction of pathways regulated by miRNAs. In recent years, small molecules have demonstrated enormous potential as drugs to regulate miRNA expression (i.e., SM-miR). A clear understanding of the mechanism of action of small molecules on the upregulation and downregulation of miRNA expression allows precise diagnosis and treatment of oncogenic pathways. However, outside of a slow and costly process of experimental determination, computational strategies to assist this on an ad hoc basis have yet to be formulated. In this work, we developed, to the best of our knowledge, the first cross-platform prediction tool, DeepsmirUD, to infer small-molecule-mediated regulatory effects on miRNA expression (i.e., upregulation or downregulation). This method is powered by 12 cutting-edge deep-learning frameworks and achieved AUC values of 0.843/0.984 and AUCPR values of 0.866/0.992 on two independent test datasets. With a complementarily constructed network inference approach based on similarity, we report a significantly improved accuracy of 0.813 in determining the regulatory effects of nearly 650 associated SM-miR relations, each formed with either novel small molecule or novel miRNA. By further integrating miRNA-cancer relationships, we established a database of potential pharmaceutical drugs from 1343 small molecules for 107 cancer diseases to understand the drug mechanisms of action and offer novel insight into drug repositioning. Furthermore, we have employed DeepsmirUD to predict the regulatory effects of a large number of high-confidence associated SM-miR relations. Taken together, our method shows promise to accelerate the development of potential miRNA targets and small molecule drugs.

Histone Deacetylase Inhibitors Restore Cancer Cell Sensitivity towards T Lymphocytes Mediated Cytotoxicity in Pancreatic Cancer

Despite medical advancements, the prognosis of pancreatic ductal adenocarcinoma (PDAC) has not improved significantly over the past 50 years. By utilising the large-scale genomic datasets available from the Australia Pancreatic Cancer Project (PACA-AU) and The Cancer Genomic Atlas Project (TCGA-PAAD), we studied the immunophenotype of PDAC in silico and identified that tumours with high cytotoxic T lymphocytes (CTL) killing activity were associated with favourable clinical outcomes. Using the STRING protein-protein interaction network analysis, the identified differentially expressed genes with low CTL killing activity were associated with TWIST/IL-6R, HDAC5, and EOMES signalling. Following Connectivity Map analysis, we identified 44 small molecules that could restore CTL sensitivity in the PDAC cells. Further high-throughput chemical library screening identified 133 inhibitors that effectively target both parental and CTL-resistant PDAC cells in vitro. Since CTL-resistant PDAC had a higher expression of histone proteins and its acetylated proteins compared to its parental cells, we further investigated the impact of histone deacetylase inhibitors (HDACi) on CTL-mediated cytotoxicity in PDAC cells in vitro, namely SW1990 and BxPC3. Further analyses revealed that givinostat and dacinostat were the two most potent HDAC inhibitors that restored CTL sensitivity in SW1990 and BxPC3 CTL-resistant cells. Through our in silico and in vitro studies, we demonstrate the novel role of HDAC inhibition in restoring CTL resistance and that combinations of HDACi with CTL may represent a promising therapeutic strategy, warranting its further detailed molecular mechanistic studies and animal studies before embarking on the clinical evaluation of these novel combined PDAC treatments.

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.

DREAM: a database of experimentally supported protein-coding RNAs and drug associations in human cancer

The Drug Response Gene Expression Associated Map, also referred as “DREAM” (http://bio-big-data.cn:8080/DREAM), is a manually curated database of experimentally supported protein-coding RNAs and drugs associations in human cancers. The current version of the DREAM documents 3048 entries about scientific literatures supported drug sensitivity or drug intervention related protein-coding RNAs from PubMed database and 195 high-throughput microarray data about drug sensitivity or drug intervention related protein-coding RNAs data from GEO database. Each entry in DREAM database contains detailed information on protein-coding RNA, drug, cancer, and other information including title, PubMed ID, journal, publish time. The DREAM database also provides some data visualization and online analysis services such as volcano plot, GO/KEGG enrichment function analysis, and novel drug discovery analysis. We hope the DREAM database should serve as a valuable resource for clinical practice and basic research, which could help researchers better understand the effects of protein-coding RNAs on drug response in human cancers.