Luhua Lai

CellChem learns cell-guided molecular representations that integrate structure and function, improving drug discovery by capturing cellular response similarities.

FuzzAletheia maps dynamic interaction patterns of IDP-ligand complexes to guide hit-to-lead optimization, achieving a 10-fold activity enhancement in a p53 ligand.

This paper develops Slab-PTM, a module for coarse-grained models to simulate post-translational modifications in intrinsically disordered proteins, showing improved accuracy in phase behavior predictions.

This paper evaluates co-folding methods for predicting ternary complexes involving molecular glues, finding AlphaFold3 performs best but relies on memorization rather than accurate modeling.

This paper introduces CYC_BUILDER, a reinforcement learning method for designing cyclic peptides that bind target proteins, demonstrating its effectiveness in generating potent inhibitors for TNF.

This paper introduces GeoSeqBuilder, a geometric deep learning framework for protein sequence design that achieves high accuracy in recovering native residues and predicting side-chain conformations.

CavitySpace identifies potential ligand binding sites in the human proteome using predicted structures, aiding in drug design and function prediction.

The paper develops a deep learning model to screen antiviral drugs against SARS-CoV-2, identifying a novel compound with promising inhibitory activity.

The paper develops a machine learning tool, PSPredictor, to predict phase separation proteins using sequence information, achieving high accuracy and outperforming existing methods.

This study develops a multi-objective metabolic model to identify cancer-associated metabolic vulnerabilities by analyzing Pareto surfaces and validating predictions with experimental data.

This study identifies HLA-DQA1 as the major independent genetic risk for ACPA-positive RA in Han population through MHC region sequencing and HLA-typing.