Chengfei Yan

A geometry-over-coevolution principle governs protein complex assembly in AlphaFold

This paper shows that geometry, not coevolution, primarily drives protein complex assembly in AlphaFold models, revealing a hierarchical mechanism for interface prediction.

Mechanism-Aware Inductive Bias Enhances Generalization in Protein-Protein Interaction Prediction

PLMDA-PPI improves protein-protein interaction prediction by integrating biophysical inductive biases, outperforming existing models in efficiency and generalization.

Protein language model embedded geometric graphs power inter-protein contact prediction

PLMGraph-Inter improves inter-protein contact prediction by integrating geometric graphs and protein language models, outperforming existing methods and enhancing docking performance.

The EN-TEx resource of multi-tissue personal epigenomes & variant-impact models

The paper introduces the EN-TEx resource, which provides personal epigenomes and variant-impact models across multiple tissues, revealing allele-specific loci and their associations with GWAS loci.

Improved inter-protein contact prediction using dimensional hybrid residual networks and protein language models

This study introduces DRN-1D2D_Inter, a deep learning method for inter-protein contact prediction that outperforms existing methods using only sequence data.

Protein Complex Structure Prediction Powered by Multiple Sequence Alignments of Interologs from Multiple Taxonomic Ranks and AlphaFold2

This study improves protein complex structure prediction by using taxonomically restricted interolog MSAs with AlphaFold2, achieving higher success rates than existing methods.

A Reproducibility Analysis-based Statistical Framework for Residue-Residue Evolutionary Coupling Detection

This paper introduces IDR-DCA, a statistical framework for detecting significant evolutionary couplings, showing improved performance over traditional methods in contact prediction and RNA structure prediction.

Improved protein contact prediction using dimensional hybrid residual networks and singularity enhanced loss function

The paper introduces a new protein contact prediction model, DRN-1D2D, which uses hybrid residual blocks and a singularity-enhanced loss function, achieving better performance than existing methods.

A cross-organism framework for supervised enhancer prediction with epigenetic pattern recognition and targeted validation

This paper develops a cross-organism framework for enhancer prediction using Drosophila STARR-seq data and supervised machine learning, validated in mouse and human with 153 enhancers.

HiC-Spector: A matrix library for spectral and reproducibility analysis of Hi-C contact maps

HiC-Spector provides a matrix library for spectral analysis and reproducibility assessment of Hi-C contact maps, distinguishing biological replicates from pseudo-replicates and different cell types.