Computational Drug Discovery Methods

This study designs non-covalent CDK7 inhibitors using computational methods to improve selectivity and safety compared to THZ1.

This study uses a Transformer model to design novel succinate dehydrogenase inhibitors against agricultural fungi, resulting in potent antifungal compounds through virtual screening and optimization.

AMODO-EO discovers and integrates new optimization criteria during drug discovery, improving multi-objective optimization by revealing hidden molecular relationships.

This study uses in silico methods to identify potential inhibitors for NDM-1 metallo-β-lactamase in K. pneumoniae, finding two sulfonamide compounds as lead candidates.

The paper uses a conditional diffusion model to design a small molecule inhibitor of secondary nucleation in alpha-synuclein aggregation, showing it effectively inhibits fibril proliferation.

DockFormer predicts protein-small molecule binding affinity and structure using a flexible docking approach, outperforming existing methods on benchmark datasets.

The paper uses machine learning to predict PPAR-γ agonist bioactivity, achieving an R2 score of 0.83 using molecular descriptors from ChEMBL data.

This paper introduces H-MoE, a hierarchical MoE model for molecular representation learning that improves generalization by incorporating scaffold-level routing and curriculum learning.

OptiRanker simulates and compares drug prioritization algorithms to rank their predictive accuracy, enabling efficient in vivo trial design for personalized medicine.

The paper introduces BioScientist Agent, a framework combining a biomedical knowledge graph, variational graph auto-encoder, reinforcement learning, and LLMs to improve drug repurposing and mechanistic reasoning.

Q-MOL improves drug discovery by accurately predicting ligand binding sites on flexible and disordered proteins, validated across diverse targets and experimental settings.

This paper predicts novel drug-drug interactions (DDIs) in women's reproductive health, identifying 2,991 new interactions and their clinical implications.

This paper identifies two small-molecule hits targeting the nNOS/CAPON interface using virtual screening and introduces computational tools for structure-based drug discovery.

The paper introduces a machine learning model that learns spatial preference maps for drug-receptor interactions to guide drug discovery.

This study uses machine learning and molecular docking to identify 14 anti-infective drugs that may be repurposed for onchocerciasis treatment, showing binding affinities similar to existing drugs.

This paper combines geometric pocket detection with large language models to prioritize protein binding pockets for drug discovery, using a curated dataset and automated pipeline.

The paper screens a compound library to identify leukemia-selective chemotypes using single-cell profiling and finds that certain rocaglate derivatives selectively induce DNA damage in leukemia cells.

This paper introduces MotifGen, a deep learning model that predicts molecular interaction motifs from receptor structures, enabling peptide binder design and small-molecule binding site prediction.

This study evaluates a hybrid model combining protein language models and graph neural networks for predicting protein-ligand binding sites, showing that structural information improves performance but complex pLMs can achieve good results without it.

This paper introduces a computational pipeline for identifying broad-spectrum antivirals by analyzing viral sequences, predicting ligand poses, and scoring protein-ligand complexes, validated on SARS-CoV-2 and MERS-CoV datasets.

This study identifies potential antiviral drug candidates against MPXV DNA polymerase and A42R protein using in-silico methods, highlighting compounds with high docking scores and stable ligand-protein complexes.

BioTransformer4.0 predicts small molecule metabolites using integrated knowledge-based and machine learning approaches, achieving high recall on benchmark datasets.

DrugPT predicts gene expression changes after drug treatment using a Transformer-based framework integrating gene and chemical representations.

This paper introduces a deterministic dynamic programming framework for molecular docking that improves reproducibility and interpretability compared to stochastic methods.

This paper proposes a framework that transfers knowledge from multi-modal genomic data to gene expression-only models to improve drug response prediction.

HTS-Oracle improves hit identification for difficult-to-drug targets by using a deep learning platform that integrates molecular embeddings and cheminformatics features, achieving an 8.4% hit rate with reduced screening costs.

This paper introduces DDMuffin, a deep learning framework for predicting protein-ligand binding affinities and mutation-induced changes, showing strong performance on benchmark datasets and offering insights for drug design.

SHARP generates drug-like molecules using reinforcement learning with a fragment-based action space, improving synthesizability and binding affinity over existing methods.

This study enhances a computational protocol for discovering covalent inhibitors of AdoMetDC by integrating quantum chemistry calculations, achieving a 70.6% hit rate in screening 17 compounds.

This paper identifies a novel druggable site on CD28 and discovers 22VS, a small molecule inhibitor that suppresses T cell activation in co-culture models, with favorable pharmacokinetic properties.