Jianpeng Ma

This paper evaluates current biomolecular structure prediction models using physics-based criteria, revealing widespread errors in capturing atomic interactions and conformational preferences.

OPUS-GO provides sequence-level and residue-level annotations for proteins and RNA sequences using a biological language model and modified MIL strategy, improving accuracy and interpretability.

OPUS-BFactor predicts normalized protein B-factors using sequence or structural information, outperforming existing methods on benchmark datasets.

OPUS-Fold3 is a gradient-based protein folding and docking framework that generates accurate 3D structures under specified constraints, outperforming pyRosetta in side-chain modeling.

OPUS-Mut improves side-chain modeling for mutation effect analysis, showing good agreement with experimental results and correlating structural shifts with functional changes.

The paper evaluates the OPUS-Mut side-chain modeling method on protein-protein interaction datasets, showing it outperforms other methods and correctly identifies native docking poses.

OPUS-Rota4 improves protein side-chain modeling by integrating deep learning predictors and constraints, achieving better accuracy than Alphafold2 on CASP14.

OPUS-X is an open-source toolkit for protein structure prediction that includes modules for torsion angles, secondary structure, solvent accessibility, contact maps, and 3D folding, outperforming existing methods in some aspects.