Unraveling druggable cancer-driving proteins and targeted drugs using artificial intelligence and multi-omics analyses.

The druggable proteome refers to proteins that can bind to small molecules with appropriate chemical affinity, inducing a favorable clinical response. Predicting druggable proteins through screening and in silico modeling is imperative for drug design. To contribute to this field, we developed an accurate predictive classifier for druggable cancer-driving proteins using amino acid composition descriptors of protein sequences and 13 machine learning linear and non-linear classifiers. The optimal classifier was achieved with the support vector machine method, utilizing 200 tri-amino acid composition descriptors. The high performance of the model is evident from an area under the receiver operating characteristics (AUROC) of 0.975 ± 0.003 and an accuracy of 0.929 ± 0.006 (threefold cross-validation). The machine learning prediction model was enhanced with multi-omics approaches, including the target-disease evidence score, the shortest pathways to cancer hallmarks, structure-based ligandability assessment, unfavorable prognostic protein analysis, and the oncogenic variome. Additionally, we performed a drug repurposing analysis to identify drugs with the highest affinity capable of targeting the best predicted proteins. As a result, we identified 79 key druggable cancer-driving proteins with the highest ligandability, and 23 of them demonstrated unfavorable prognostic significance across 16 TCGA PanCancer types: CDKN2A, BCL10, ACVR1, CASP8, JAG1, TSC1, NBN, PREX2, PPP2R1A, DNM2, VAV1, ASXL1, TPR, HRAS, BUB1B, ATG7, MARK3, SETD2, CCNE1, MUTYH, CDKN2C, RB1, and SMARCA4. Moreover, we prioritized 11 clinically relevant drugs targeting these proteins. This strategy effectively predicts and prioritizes biomarkers, therapeutic targets, and drugs for in-depth studies in clinical trials. Scripts are available at https://github.com/muntisa/machine-learning-for-druggable-proteins .

Repurposed AT9283 triggers anti-tumoral effects by targeting MKK3 oncogenic functions in Colorectal Cancer.

BACKGROUND: Colorectal cancer (CRC) is the third most common type of cancer and the second leading cause of cancer-related deaths worldwide, with a survival rate near to 10% when diagnosed at an advanced stage. Hence, the identification of new molecular targets to design more selective and efficient therapies is urgently required. The Mitogen activated protein kinase kinase 3 (MKK3) is a dual-specificity threonine/tyrosine protein kinase that, activated in response to cellular stress and inflammatory stimuli, regulates a plethora of biological processes. Previous studies revealed novel MKK3 roles in supporting tumor malignancy, as its depletion induces autophagy and cell death in cancer lines of different tumor types, including CRC. Therefore, MKK3 may represent an interesting new therapeutic target in advanced CRC, however selective MKK3 inhibitors are currently not available. METHODS: The study involved transcriptomic based drug repurposing approach and confirmatory assays with CRC lines, primary colonocytes and a subset of CRC patient-derived organoids (PDO). Investigations in vitro and in vivo were addressed. RESULTS: The repurposing approach identified the multitargeted kinase inhibitor AT9283 as a putative compound with MKK3 depletion-mimicking activities. Indeed, AT9283 drops phospho- and total-MKK3 protein levels in tested CRC models. Likely the MKK3 silencing, AT9283 treatment: i) inhibited cell proliferation promoting autophagy and cell death in tested CRC lines and PDOs; ii) resulted well-tolerated by CCD-18Co colonocytes; iii) reduced cancer cell motility inhibiting CRC cell migration and invasion; iv) inhibited COLO205 xenograft tumor growth. Mechanistically, AT9283 abrogated MKK3 protein levels mainly through the inhibition of aurora kinase A (AURKA), impacting on MKK3/AURKA protein-protein interaction and protein stability therefore uncovering the relevance of MKK3/AURKA crosstalk in sustaining CRC malignancy in vitro and in vivo. CONCLUSION: Overall, we demonstrated that the anti-tumoral effects triggered by AT9283 treatment recapitulated the MKK3 depletion effects in all tested CRC models in vitro and in vivo, suggesting that AT9283 is a repurposed drug. According to its good tolerance when tested with primary colonocytes (CCD-18CO), AT9283 is a promising drug for the development of novel therapeutic strategies to target MKK3 oncogenic functions in late-stage and metastatic CRC patients.

Multi-omics analysis identifies repurposing bortezomib in the treatment of kidney-, nervous system-, and hematological cancers.

Repurposing of FDA-approved drugs is a quick and cost-effective alternative to de novo drug development. Here, we identify genes involved in bortezomib sensitivity, predict cancer types that may benefit from treatment with bortezomib, and evaluate the mechanism-of-action of bortezomib in breast cancer (BT-474 and ZR-75-30), melanoma (A-375), and glioblastoma (A-172) cells in vitro. Cancer cell lines derived from cancers of the blood, kidney, nervous system, and skin were found to be significantly more sensitive to bortezomib than other organ systems. The in vitro studies confirmed that although bortezomib effectively inhibited the ß5 catalytic site in all four cell lines, cell cycle arrest was only induced in G2/M phase and apoptosis in A-375 and A-172 after 24h. The genomic and transcriptomic analyses identified 33 genes (e.g. ALDH18A1, ATAD2) associated with bortezomib resistance. Taken together, we identified biomarkers predictive of bortezomib sensitivity and cancer types that might benefit from treatment with bortezomib.

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 ß5 subunit, our in vitro study revealed that these compounds inhibited the ß1, ß2, and ß5 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.

Exploring the molecular mechanisms and shared potential drugs between rheumatoid arthritis and arthrofibrosis based on large language model and synovial microenvironment analysis.

Rheumatoid arthritis (RA) and arthrofibrosis (AF) are both chronic synovial hyperplasia diseases that result in joint stiffness and contractures. They shared similar symptoms and many common features in pathogenesis. Our study aims to perform a comprehensive analysis between RA and AF and identify novel drugs for clinical use. Based on the text mining approaches, we performed a correlation analysis of 12 common joint diseases including arthrofibrosis, gouty arthritis, infectious arthritis, juvenile idiopathic arthritis, osteoarthritis, post infectious arthropathies, post traumatic osteoarthritis, psoriatic arthritis, reactive arthritis, rheumatoid arthritis, septic arthritis, and transient arthritis. 5 bulk sequencing datasets and 4 single-cell sequencing datasets of RA and AF were integrated and analyzed. A novel drug repositioning method was found for drug screening, and text mining approaches were used to verify the identified drugs. RA and AF performed the highest gene similarity (0.77) and functional ontology similarity (0.84) among all 12 joint diseases. We figured out that they share the same key pathogenic cell including CD34 + sublining fibroblasts (CD34-SLF) and DKK3 + sublining fibroblasts (DKK3-SLF). Potential therapeutic target database (PTTD) was established with the differential expressed genes (DEGs) of these key pathogenic cells. Based on the PTTD, 15 potential drugs for AF and 16 potential drugs for RA were identified. This work provides a new perspective on AF and RA study which enhances our understanding of their pathogenesis. It also shed light on their underlying mechanism and open new avenues for drug repositioning studies.

Repurposing Anidulafungin for Alzheimer's Disease via Fragment-Based Drug Discovery.

The misfolding and aggregation of beta-amyloid (Aß) peptides have been implicated as key pathogenic events in the early stages of Alzheimer's disease (AD). Inhibiting Aß aggregation represents a potential disease-modifying therapeutic approach to AD treatment. Previous studies have identified various molecules that inhibit Aß aggregation, some of which share common chemical substructures (fragments) that may be key to their inhibitory activity. Employing fragment-based drug discovery (FBDD) methods may facilitate the identification of these fragments, which can subsequently be used to screen new inhibitors and provide leads for further drug development. In this study, we used an in silico FBDD approach to identify 17 fragment clusters that are significantly enriched among Aß aggregation inhibitors. These fragments were then used to screen anti-infective agents, a promising drug class for repurposing against amyloid aggregation. This screening process identified 16 anti-infective drugs, 5 of which were chosen for further investigation. Among the 5 candidates, anidulafungin, an antifungal compound, showed high efficacy in inhibiting Aß aggregation in vitro. Kinetic analysis revealed that anidulafungin selectively blocks the primary nucleation step of Aß aggregation, substantially delaying Aß fibril formation. Cell viability assays demonstrated that anidulafungin can reduce the toxicity of oligomeric Aß on BV2 microglia cells. Molecular docking simulations predicted that anidulafungin interacted with various Aß species, including monomers, oligomers, and fibrils, potentially explaining its activity against Aß aggregation and toxicity. This study suggests that anidulafungin is a potential drug to be repurposed for AD, and FBDD is a promising approach for discovering drugs to combat Aß aggregation.

Mechanistic insights into antifungal potential of Alexidine dihydrochloride and hexachlorophene in Candida albicans: a drug repurposing approach.

Candida albicans has been listed in the critical priority group by the WHO in 2022 depending upon its contribution in invasive candidiasis and increased resistance to conventional drugs. Drug repurposing offers an efficient, rapid, and cost-effective solution to develop alternative therapeutics against pathogenic microbes. Alexidine dihydrochloride (AXD) and hexachlorophene (HCP) are FDA approved anti-cancer and anti-septic drugs, respectively. In this study, we have shown antifungal properties of AXD and HCP against the wild type (reference strain) and clinical isolates of C. albicans. The minimum inhibitory concentrations (MIC50) of AXD and HCP against C. albicans ranged between 0.34 and 0.69 µM and 19.66-24.58 µM, respectively. The biofilm inhibitory and eradication concentration of AXD was reported comparatively lower than that of HCP for the strains used in the study. Further investigations were performed to understand the antifungal mode of action of AXD and HCP by studying virulence features like cell surface hydrophobicity, adhesion, and yeast to hyphae transition, were also reduced upon exposure to both the drugs. Ergosterol content in cell membrane of the wild type strain was upregulated on exposure to AXD and HCP both. Biochemical analyses of the exposed biofilm indicated reduced contents of carbohydrate, protein, and e-DNA in the extracellular matrix of the biofilm when compared to the untreated control biofilm. AXD exposure downregulated activity of tissue invading enzyme, phospholipase in the reference strain. In wild type strain, ROS level, and activities of antioxidant enzymes were found elevated upon exposure to both drugs. FESEM analysis of the drug treated biofilms revealed degraded biofilm. This study has indicated mode of action of antifungal potential of alexidine dihydrochloride and hexachlorophene in C. albicans.

Drug repositioning based on residual attention network and free multiscale adversarial training.

BACKGROUND: Conducting traditional wet experiments to guide drug development is an expensive, time-consuming and risky process. Analyzing drug function and repositioning plays a key role in identifying new therapeutic potential of approved drugs and discovering therapeutic approaches for untreated diseases. Exploring drug-disease associations has far-reaching implications for identifying disease pathogenesis and treatment. However, reliable detection of drug-disease relationships via traditional methods is costly and slow. Therefore, investigations into computational methods for predicting drug-disease associations are currently needed. RESULTS: This paper presents a novel drug-disease association prediction method, RAFGAE. First, RAFGAE integrates known associations between diseases and drugs into a bipartite network. Second, RAFGAE designs the Re_GAT framework, which includes multilayer graph attention networks (GATs) and two residual networks. The multilayer GATs are utilized for learning the node embeddings, which is achieved by aggregating information from multihop neighbors. The two residual networks are used to alleviate the deep network oversmoothing problem, and an attention mechanism is introduced to combine the node embeddings from different attention layers. Third, two graph autoencoders (GAEs) with collaborative training are constructed to simulate label propagation to predict potential associations. On this basis, free multiscale adversarial training (FMAT) is introduced. FMAT enhances node feature quality through small gradient adversarial perturbation iterations, improving the prediction performance. Finally, tenfold cross-validations on two benchmark datasets show that RAFGAE outperforms current methods. In addition, case studies have confirmed that RAFGAE can detect novel drug-disease associations. CONCLUSIONS: The comprehensive experimental results validate the utility and accuracy of RAFGAE. We believe that this method may serve as an excellent predictor for identifying unobserved disease-drug associations.

Computational analysis of pathogen-host interactome for fast and low-risk in-silico drug repurposing in emerging viral threats like Mpox.

Monkeypox (Mpox), a zoonotic illness triggered by the monkeypox virus (MPXV), poses a significant threat since it may be transmitted and has no cure. This work introduces a computational method to predict Protein-Protein Interactions (PPIs) during MPXV infection. The objective is to discover prospective drug targets and repurpose current potential Food and Drug Administration (FDA) drugs for therapeutic purposes. In this work, ensemble features, comprising 2-5 node graphlet attributes and protein composition-based features are utilized for Deep Learning (DL) models to predict PPIs. The technique that is used here demonstrated an excellent prediction performance for PPI on both the Human Integrated Protein-Protein Interaction Reference (HIPPIE) and MPXV-Human PPI datasets. In addition, the human protein targets for MPXV have been identified accurately along with the detection of possible therapeutic targets. Furthermore, the validation process included conducting docking research studies on potential FDA drugs like Nicotinamide Adenine Dinucleotide and Hydrogen (NADH), Fostamatinib, Glutamic acid, Cannabidiol, Copper, and Zinc in DrugBank identified via research on drug repurposing and the Drug Consensus Score (DCS) for MPXV. This has been achieved by employing the primary crystal structures of MPXV, which are now accessible. The docking study is also supported by Molecular Dynamics (MD) simulation. The results of our study emphasize the effectiveness of using ensemble feature-based PPI prediction to understand the molecular processes involved in viral infection and to aid in the development of repurposed drugs for emerging infectious diseases such as, but not limited to, Mpox. The source code and link to data used in this work is available at: https://github.com/CMATERJU-BIOINFO/In-Silico-Drug-Repurposing-Methodology-To-Suggest-Therapies-For-Emerging-Threats-like-Mpox .

Recent Development of Fluoroquinolone Derivatives as Anticancer Agents.

Cancer is the second leading cause of death in the world following cardiovascular disease. Its treatment, including radiation therapy and surgical removal of the tumour, is based on pharmacotherapy, which prompts a constant search for new and more effective drugs. There are high costs associated with designing, synthesising, and marketing new substances. Drug repositioning is an attractive solution. Fluoroquinolones make up a group of synthetic antibiotics with a broad spectrum of activity in bacterial diseases. Moreover, those compounds are of particular interest to researchers as a result of reports of their antiproliferative effects on the cells of the most lethal cancers. This article presents the current progress in the development of new fluoroquinolone derivatives with potential anticancer and cytotoxic activity, as well as structure-activity relationships, along with possible directions for further development.

Chemical Coaxing of Mesenchymal Stromal Cells by Drug Repositioning for Nestin Induction.

Mesenchymal stromal cells (MSCs) display heterogeneity in origin and functional role in tissue homeostasis. Subsets of MSCs derived from the neural crest express nestin and serve as niches in bone marrow, but the possibility of coaxing MSCs into nestin-expresing cells for enhanced supportive activity is unclear. In this study, as an approach to the chemical coaxing of MSC functions, we screened libraries of clinically approved chemicals to identify compounds capable of inducing nestin expression in MSCs. Out of 2000 clinical compounds, we chose vorinostat as a candidate to coax the MSCs into neural crest-like fates. When treated with vorinostat, MSCs exhibited a significant increase in the expression of genes involved in the pluripotency and epithelial-mesenchymal transition (EMT), as well as nestin and CD146, the markers for pericytes. In addition, these nestin-induced MSCs exhibited enhanced differentiation towards neuronal cells with the upregulation of neurogenic markers, including SRY-box transcription factor 2 (Sox2), SRY-box transcription factor 10 (Sox10) and microtubule associated protein 2 (Map2) in addition to nestin. Moreover, the coaxed MSCs exhibited enhanced supporting activity for hematopoietic progenitors without supporting leukemia cells. These results demonstrate the feasibility of the drug repositioning of MSCs to induce neural crest-like properties through the chemical coaxing of cell fates.

From Proteome to Potential Drugs: Integration of Subtractive Proteomics and Ensemble Docking for Drug Repurposing against Pseudomonas aeruginosa RND Superfamily Proteins.

Pseudomonas aeruginosa (P. aeruginosa) poses a significant threat as a nosocomial pathogen due to its robust resistance mechanisms and virulence factors. This study integrates subtractive proteomics and ensemble docking to identify and characterize essential proteins in P. aeruginosa, aiming to discover therapeutic targets and repurpose commercial existing drugs. Using subtractive proteomics, we refined the dataset to discard redundant proteins and minimize potential cross-interactions with human proteins and the microbiome proteins. We identified 12 key proteins, including a histidine kinase and members of the RND efflux pump family, known for their roles in antibiotic resistance, virulence, and antigenicity. Predictive modeling of the three-dimensional structures of these RND proteins and subsequent molecular ensemble-docking simulations led to the identification of MK-3207, R-428, and Suramin as promising inhibitor candidates. These compounds demonstrated high binding affinities and effective inhibition across multiple metrics. Further refinement using non-covalent interaction index methods provided deeper insights into the electronic effects in protein-ligand interactions, with Suramin exhibiting superior binding energies, suggesting its broad-spectrum inhibitory potential. Our findings confirm the critical role of RND efflux pumps in antibiotic resistance and suggest that MK-3207, R-428, and Suramin could be effectively repurposed to target these proteins. This approach highlights the potential of drug repurposing as a viable strategy to combat P. aeruginosa infections.

PBPK-led assessment of antimalarial drugs as candidates for Covid-19: Simulating concentrations at the site of action to inform repurposing strategies.

The urgent need for safe, efficacious, and accessible drug treatments to treat coronavirus disease 2019 (COVID-19) prompted a global effort to evaluate drug repurposing opportunities. Pyronaridine and amodiaquine are both components of approved antimalarials with in vitro activity against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). In vitro activity does not always translate to clinical efficacy across a therapeutic dose range. This study applied available, verified, physiologically based pharmacokinetic (PBPK) models for pyronaridine, amodiaquine, and its active metabolite N-desethylamodiaquine (DEAQ) to predict drug concentrations in lung tissue relative to plasma or blood in the default healthy virtual population. Lung exposures were compared to published data across the reported range of in vitro EC50 values against SARS-CoV-2. In the multicompartment permeability-limited PBPK model, the predicted total Cmax in lung mass for pyronaridine was 34.2 µM on Day 3, 30.5-fold greater than in blood (1.12 µM) and for amodiaquine was 0.530 µM, 8.83-fold greater than in plasma (0.060 µM). In the perfusion-limited PBPK model, the DEAQ predicted total Cmax on Day 3 in lung mass (30.2 µM) was 21.4-fold greater than for plasma (1.41 µM). Based on the available in vitro data, predicted drug concentrations in lung tissue for pyronaridine and DEAQ, but not amodiaquine, appeared sufficient to inhibit SARS-CoV-2 replication. Simulations indicated standard dosing regimens of pyronaridine-artesunate and artesunate-amodiaquine have potential to treat COVID-19. These findings informed repurposing strategies to select the most relevant compounds for clinical investigation in COVID-19. Clinical data for model verification may become available from ongoing clinical studies.

In-silico discovery of common molecular signatures for which SARS-CoV-2 infections and lung diseases stimulate each other, and drug repurposing.

COVID-19 caused by SARS-CoV-2 is a global health issue. It is yet a severe risk factor to the patients, who are also suffering from one or more chronic diseases including different lung diseases. In this study, we explored common molecular signatures for which SARS-CoV-2 infections and different lung diseases stimulate each other, and associated candidate drug molecules. We identified both SARS-CoV-2 infections and different lung diseases (Asthma, Tuberculosis, Cystic Fibrosis, Pneumonia, Emphysema, Bronchitis, IPF, ILD, and COPD) causing top-ranked 11 shared genes (STAT1, TLR4, CXCL10, CCL2, JUN, DDX58, IRF7, ICAM1, MX2, IRF9 and ISG15) as the hub of the shared differentially expressed genes (hub-sDEGs). The gene ontology (GO) and pathway enrichment analyses of hub-sDEGs revealed some crucial common pathogenetic processes of SARS-CoV-2 infections and different lung diseases. The regulatory network analysis of hub-sDEGs detected top-ranked 6 TFs proteins and 6 micro RNAs as the key transcriptional and post-transcriptional regulatory factors of hub-sDEGs, respectively. Then we proposed hub-sDEGs guided top-ranked three repurposable drug molecules (Entrectinib, Imatinib, and Nilotinib), for the treatment against COVID-19 with different lung diseases. This recommendation is based on the results obtained from molecular docking analysis using the AutoDock Vina and GLIDE module of Schrödinger. The selected drug molecules were optimized through density functional theory (DFT) and observing their good chemical stability. Finally, we explored the binding stability of the highest-ranked receptor protein RELA with top-ordered three drugs (Entrectinib, Imatinib, and Nilotinib) through 100 ns molecular dynamic (MD) simulations with YASARA and Desmond module of Schrödinger and observed their consistent performance. Therefore, the findings of this study might be useful resources for the diagnosis and therapies of COVID-19 patients who are also suffering from one or more lung diseases.

Repurposing Anticancer Drugs Targeting the MAPK/ERK Signaling Pathway for the Treatment of Respiratory Virus Infections.

Respiratory virus infections remain a significant challenge to human health and the social economy. The symptoms range from mild rhinitis and nasal congestion to severe lower respiratory tract dysfunction and even mortality. The efficacy of therapeutic drugs targeting respiratory viruses varies, depending upon infection time and the drug resistance engendered by a high frequency of viral genome mutations, necessitating the development of new strategies. The MAPK/ERK pathway that was well delineated in the 1980s represents a classical signaling cascade, essential for cell proliferation, survival, and differentiation. Since this pathway is constitutively activated in many cancers by oncogenes, several drugs inhibiting Raf/MEK/ERK have been developed and currently used in anticancer treatment. Two decades ago, it was reported that viruses such as HIV and influenza viruses could exploit the host cellular MAPK/ERK pathway for their replication. Thus, it would be feasible to repurpose this category of the pathway inhibitors for the treatment of respiratory viral infections. The advantage is that the host genes are not easy to mutate such that the drug resistance rarely occurs during short-period treatment of viruses. Therefore, in this review we will summarize the research progress on the role of the MAPK/ERK pathway in respiratory virus amplification and discuss the potential of the pathway inhibitors (MEK inhibitors) in the treatment of respiratory viral infections.

Repurposing FDA-approved drugs to treat chemical weapon toxicities: Interactive case studies for trainees.

The risk of a terrorist attack in the United States has created challenges on how to effectively treat toxicities that result from exposure to chemical weapons. To address this concern, the United States has organized a trans-agency initiative across academia, government, and industry to identify drugs to treat tissue injury resulting from exposure to chemical threat agents. We sought to develop and evaluate an interactive educational session that provides hands-on instruction on how to repurpose FDA-approved drugs as therapeutics to treat toxicity from exposure to chemical weapons. As part of the Rutgers Summer Undergraduate Research Fellowship program, 23 undergraduate students participated in a 2-h session that included: (1) an overview of chemical weapon toxicities, (2) a primer on pharmacology principles, and (3) an interactive session where groups of students were provided lists of FDA-approved drugs to evaluate potential mechanisms of action and suitability as countermeasures for four chemical weapon case scenarios. The interactive session culminated in a competition for the best grant "sales pitch." From this interactive training, students improved their understanding of (1) the ability of chemical weapons to cause long-term toxicities, (2) impact of route of administration and exposure scenario on drug efficacy, and (3) re-purposing FDA-approved drugs to treat disease from chemical weapon exposure. These findings demonstrated that an interactive training exercise can provide students with new insights into drug development for chemical threat agent toxicities.

An experimentally validated approach to automated biological evidence generation in drug discovery using knowledge graphs.

Explaining predictions for drug repositioning with biological knowledge graphs is a challenging problem. Graph completion methods using symbolic reasoning predict drug treatments and associated rules to generate evidence representing the therapeutic basis of the drug. Yet the vast amounts of generated paths that are biologically irrelevant or not mechanistically meaningful within the context of disease biology can limit utility. We use a reinforcement learning based knowledge graph completion model combined with an automatic filtering approach that produces the most relevant rules and biological paths explaining the predicted drug's therapeutic connection to the disease. In this work we validate the approach against preclinical experimental data for Fragile X syndrome demonstrating strong correlation between automatically extracted paths and experimentally derived transcriptional changes of selected genes and pathways of drug predictions Sulindac and Ibudilast. Additionally, we show it reduces the number of generated paths in two case studies, 85% for Cystic fibrosis and 95% for Parkinson's disease.

RepurposeDrugs: an interactive web-portal and predictive platform for repurposing mono- and combination therapies.

RepurposeDrugs (https://repurposedrugs.org/) is a comprehensive web-portal that combines a unique drug indication database with a machine learning (ML) predictor to discover new drug-indication associations for approved as well as investigational mono and combination therapies. The platform provides detailed information on treatment status, disease indications and clinical trials across 25 indication categories, including neoplasms and cardiovascular conditions. The current version comprises 4314 compounds (approved, terminated or investigational) and 161 drug combinations linked to 1756 indications/conditions, totaling 28 148 drug-disease pairs. By leveraging data on both approved and failed indications, RepurposeDrugs provides ML-based predictions for the approval potential of new drug-disease indications, both for mono- and combinatorial therapies, demonstrating high predictive accuracy in cross-validation. The validity of the ML predictor is validated through a number of real-world case studies, demonstrating its predictive power to accurately identify repurposing candidates with a high likelihood of future approval. To our knowledge, RepurposeDrugs web-portal is the first integrative database and ML-based predictor for interactive exploration and prediction of both single-drug and combination approval likelihood across indications. Given its broad coverage of indication areas and therapeutic options, we expect it accelerates many future drug repurposing projects.

DeepDRA: Drug repurposing using multi-omics data integration with autoencoders.

Cancer treatment has become one of the biggest challenges in the world today. Different treatments are used against cancer; drug-based treatments have shown better results. On the other hand, designing new drugs for cancer is costly and time-consuming. Some computational methods, such as machine learning and deep learning, have been suggested to solve these challenges using drug repurposing. Despite the promise of classical machine-learning methods in repurposing cancer drugs and predicting responses, deep-learning methods performed better. This study aims to develop a deep-learning model that predicts cancer drug response based on multi-omics data, drug descriptors, and drug fingerprints and facilitates the repurposing of drugs based on those responses. To reduce multi-omics data's dimensionality, we use autoencoders. As a multi-task learning model, autoencoders are connected to MLPs. We extensively tested our model using three primary datasets: GDSC, CTRP, and CCLE to determine its efficacy. In multiple experiments, our model consistently outperforms existing state-of-the-art methods. Compared to state-of-the-art models, our model achieves an impressive AUPRC of 0.99. Furthermore, in a cross-dataset evaluation, where the model is trained on GDSC and tested on CCLE, it surpasses the performance of three previous works, achieving an AUPRC of 0.72. In conclusion, we presented a deep learning model that outperforms the current state-of-the-art regarding generalization. Using this model, we could assess drug responses and explore drug repurposing, leading to the discovery of novel cancer drugs. Our study highlights the potential for advanced deep learning to advance cancer therapeutic precision.