Pratyush Tiwary

Topology, Energy Landscapes and Folding Kinetics of RNA Hairpins

This study develops a computational framework to model RNA conformational ensembles and their dynamics, showing temperature's effect on folding landscapes and agreement with experimental data.

Early-Enrichment Hit Discovery via Reversible-Work c(t) Estimation in Metadynamics (CTMD)

This paper introduces CTMD, a physics-based method for early enrichment in virtual screening that outperforms approximate scoring functions and AI-based co-folding methods while being fast and transferable.

MAHLER: Integrating Metadynamics and Inverse Folding to Predict Antibody-Antigen Kinetics

MAHLER predicts antibody-antigen dissociation kinetics using a hybrid machine learning and physics approach, achieving screening-grade accuracy and fast computation.

Detecting and quantifying overparametrization in RNA language models with REDIAL

This paper introduces REDIAL, an unsupervised framework to detect overparametrization in RNA language models, revealing inefficiencies and design flaws in current models.

Ab initio prediction of RNA structure ensembles with RNAnneal

RNAnneal predicts RNA structural ensembles using deep learning and physics-based modeling, outperforming existing methods in classifying experimentally resolved conformations.

Using Time Dependent Rate Analysis to Evaluate the Quality of Machine Learned Reaction Coordinates for Biasing and Computing Kinetics

This study evaluates the quality of machine learned reaction coordinates using the gamma metric and shows an inverse correlation with accelerated times.

Machine learning-augmented molecular dynamics simulations (MD) reveal insights into the disconnect between affinity and activation of ZTPriboswitch ligands

The paper uses machine learning-augmented MD simulations to study ZTP riboswitch ligand interactions, finding that ligand on-rates, not binding affinity, determine activation potency.

Calculating Protein-Ligand Residence Times Through State Predictive Information Bottleneck based Enhanced Sampling

This paper introduces a method to calculate protein-ligand residence times across wide timescales using deep learning and enhanced sampling.