P. Tiwary

Graph Attention Site Prediction (GrASP): Identifying Druggable Binding Sites Using Graph Neural Networks with Attention

GrASP identifies druggable protein binding sites using graph neural networks with attention, achieving state-of-the-art performance and high precision.

Learning high-dimensional reaction coordinates of fast-folding proteins using State Predictive Information Bottleneck and Bias Exchange Metadynamics

This study uses AI and enhanced sampling to accelerate protein unfolding simulations by learning high-dimensional reaction coordinates from poor data.

AlphaFold2-RAVE: From sequence to Boltzmann ensemble

AlphaFold2-RAVE generates thermodynamically ranked conformational ensembles using AlphaFold2 structures as initializations for AI-enhanced molecular dynamics simulations.

Interrogating RNA-small molecule interactions with structure probing and AI augmented-molecular simulations

The paper uses AI-augmented molecular dynamics simulations to study RNA-ligand interactions, validating predictions of mutations affecting binding affinities in riboswitch systems.

Protein flexibility and dissociation pathway differentiation can explain onset of resistance mutations in kinases

This paper uses molecular dynamics simulations to explain how a single point mutation in Abl kinase leads to different dissociation pathways and reduced drug residence time for Imatinib.

Making high-dimensional molecular distribution functions tractable through Belief Propagation on Factor Graphs

This paper develops a belief propagation method on factor graphs to approximate high-dimensional molecular distribution functions, enabling faster sampling and analysis of conformational dynamics in peptides.

Confronting pitfalls of AI-augmented molecular dynamics using statistical physics

This paper introduces an automated algorithm to improve AI-augmented molecular dynamics by maximizing time-scale separation, tested on three benchmark problems.

Discovering loop conformational flexibility in T4 lysozyme mutants through Artificial Intelligence aided Molecular Dynamics

This study uses AI-aided molecular dynamics to analyze conformational flexibility in T4 lysozyme mutants, revealing how sequence changes affect loop stability and dynamics.

Automatic mutual information noise omission (AMINO): generating order parameters for molecular systems

AMINO generates order parameters for molecular systems using information theory, enabling efficient enhanced sampling and accurate binding affinity calculations.

Can one trust kinetic and thermodynamic observables from biased metadynamics simulations: detailed quantitative benchmarks on millimolar drug fragment dissociation

This study benchmarks infrequent metadynamics against unbiased MD simulations to assess the reliability of kinetic and thermodynamic observables for ligand dissociation from FKBP.

Past-future information bottleneck framework for simultaneously sampling biomolecular reaction coordinate, thermodynamics and kinetics

This paper introduces a framework to sample biomolecular reaction coordinates, thermodynamics, and kinetics using predictive information bottleneck methods, demonstrating it on lysozyme benzene dissociation.

Multi-dimensional spectral gap optimization of order parameters (SGOOP) through conditional probability factorization

This paper extends the SGOOP method to construct multi-dimensional reaction coordinates by using conditional probability factorization, demonstrating its effectiveness in simulating complex molecular systems.