Nishi Maru

A library for measuring chirality at the level of individual atomic environments, a continuous chirality measure with a Rust backend and a Python API. Co-first-authored with Aster, targeting Chemical Communications.

Studying the mechanistic details of Rh-catalysed asymmetric reaction reactions. This is in collaboration with Prof Stephen P. Fletcher.

Developing a Δ-machine learning model to predict DFT properties for NHCs, using xTB descriptors.

Building ML models for the prediction of reaction outcomes in NHC-catalysed reactions.

Built ML models for yield prediction.

Mining chemical structures of odour-causing and order-removing compounds from patent documents.

Developed force field models using classical methods, using General Amber Force Field (GAFF) and Tailor-Made Force Fields (TMFF), for an accurate representation of molecular systems in molecular dynamics simulations.

Conducted docking studies using classical method - AutoDock Vina and a diffusion-based model - DiffDock Built a pipeline for filtering molecules using RDKit - based on Lipinski's Rule of 5 + a few medicinal chemistry aspects

Studied Ni-catalysed cross-electrophile reactions, its mechanism and pathway and provided insights in optimising the conditions and product selectivity. Implemented xTB to accelerate DFT jobs, reducing computational time while maintaining reasonable accuracy.

Designed molecular descriptors using xTB and tested various Machine learning algorithms to predict solubility. Tested the ML metrics with models developed using DFT-based molecular descriptors.