Ben Geoffrey A S

PATHNet improves protein-ligand interaction prediction by combining physics-inspired modeling with neural networks, achieving high accuracy across multiple targets.

DMPKformer predicts ADMET properties using a multimodal deep learning framework that integrates molecular representations, achieving competitive performance and interpretability.

This paper introduces MOLDE, an in-silico method for designing molecular glues to target undruggable proteins by leveraging known ternary complexes.

This paper develops a Bayesian weighted scheme to predict SARS-CoV lineage evolution and demonstrates its retrospective application to anticipate SARS-CoV-2 emergence from SARS-CoV-1.

This paper introduces a computational method for in-silico PROTAC design and evaluation, demonstrating its ability to predict PROTAC binding modes and thermodynamics for known and novel target systems.

This paper establishes an in-silico rationale for molecular glue design by analyzing a known case and discusses its application in designing novel molecular glues using AI techniques.

This paper presents an in-silico approach for PROTAC design that uses AI and molecular dynamics to predict ternary complex stability and quantify PROTAC performance metrics.

The paper trains a quantum-classical hybrid neural network on the CASP dataset to predict protein structures and evaluates its performance using contact maps.

The paper presents a Python program that screens over a billion compounds from PubChem using machine learning to generate drug leads for Dengue and West Nile viruses, followed by automated In Silico modelling of their interactions with methyltransferase targets.

The paper introduces a programmatic tool that automates drug discovery for Hepatitis C virus using machine learning and molecular dynamics simulations to identify potential drug candidates.