Computational Drug Discovery Methods

This study uses molecular docking to evaluate the anti-HCC potential of compounds from Celtis aetnensis against three cancer-related receptors.

The paper predicts pIC50 values for antiviral peptides against Flaviviridae targets using proteochemometric modeling, achieving R2 of 0.85 and MAPE of 8.44%.

This study predicts antiviral-HPV protein interactions using machine learning to identify potential drug candidates for anti-HPV pharmacotherapy.

The paper develops machine learning models to predict active molecules against breast cancer cells, achieving high accuracy across different cell subtypes.

InDeep predicts functional binding sites in proteins for drug design on PPIs, outperforming existing tools on both PPI and conventional targets.

This paper introduces MolCLE, a molecular graph contrastive learning framework with parameterized explainable augmentations, which outperforms baselines on real-world datasets with fewer parameters.

This paper introduces a deep learning approach to design drug-like compounds by generating molecules that mimic interaction signatures of existing drugs, showing significant behavioral similarity in 16/20 cases.

The paper develops TRANSACT, a transfer learning framework for predicting clinical drug response, outperforming existing methods on multiple datasets.

This study uses UUMD simulations to reconstruct the binding pathway of Indinavir to HTLV-1 protease, identifying key residues and intermediate states involved in the interaction.

DRUIDom predicts compound-target interactions using protein domain modularity, identifying 27,032 high-confidence associations and validating them experimentally on LIM kinases.

The paper identifies an unexpected similarity between HIV-1 reverse transcriptase and tumor necrosis factor binding sites using a new computational method, confirmed by experimental assays.

The paper uses a random forest classifier to predict which drugs are detectable on skin using metabolomics, finding that nervous system drugs are more likely to be detected.

This study proposes a deep learning model with an attention mechanism to predict protein-ligand binding affinity, showing improved performance on benchmark datasets.

This study identifies neuromodulators from Curcuma amada targeting Acetylcholinesterase using in silico methods, showing drug-likeness and potential therapeutic effects for neurodegenerative disorders.

This study identifies potential bioactive compounds that inhibit metastasis tumor antigens (MTAs) and suggests therapeutic options to prevent cancer progression.

The paper evaluates the conformational space of the estrogen receptor alpha (ER) using crystal structures, molecular dynamics, and virtual screening to inform drug design.

PharmacoDB 2.0 enhances pharmacogenomics analysis by integrating new datasets, implementing FAIR pipelines, and improving drug response analysis tools.

DINC-COVID provides a webserver for ensemble docking of SARS-CoV-2 proteins with flexible receptors and ligands, generating and scoring plausible binding modes.

The paper introduces a deep learning framework that generates 3D coordinates of drug-like molecules with desired scaffolds, showing it can create valid, novel, and synthesizable molecules with drug-like properties.

The paper predicts biochemical functions and potential drug targets for 2103 hypothetical proteins in Pseudomonas aeruginosa using homology modelling, identifying 11 promising targets for antibiotic development.

This study compares the unbinding pathways of Indinavir from HIV and HTLV-1 proteases using Supervised Molecular Dynamics to explain its weak inhibitory effect on HTLV-1.

DLAB uses deep learning to improve antibody-antigen docking and virtual screening by predicting binding and ranking poses for antibody drug candidates.

This study identifies four bioflavonoids from Azadirachta indica that inhibit Japanese encephalitis virus RdRp through molecular docking and dynamics simulations.

DeepDDS uses a graph neural network with attention to predict synergistic drug combinations, outperforming existing methods on benchmark and independent test data.

This study evaluates the reproducibility of the Connectivity Map (CMap) for drug repositioning and finds limited reliability between CMap versions and low differential expression consistency.

This paper introduces a method to calculate ligand strain energy using torsional populations from the Cambridge Structural Database, improving docking hit rates by filtering strained decoys.

The paper identifies novel small molecule inhibitors for drug-resistant tuberculosis proteins using computational methods and docking analysis.

This paper reviews and compares performance evaluation measures for drug repurposing technologies using the CANDO platform to improve cross-platform comparability and optimize drug discovery.

This paper uses probabilistic matrix factorization to predict synergistic drug combinations from partial data, showing high accuracy and proposing an efficient experimental design.

The paper develops a deep learning model called DEERS to predict kinase inhibitor efficacy in cancer cell lines, achieving high accuracy and providing interpretable insights into drug response mechanisms.