Xiangxiang Zeng

HapFold improves haplotype reconstruction for diploid genomes by combining graph-based and sequence-based scaffolding methods, achieving high accuracy and contiguity.

KGIPA improves prediction of peptide-protein non-covalent interactions by integrating biological knowledge and multimodal features, outperforming existing methods in binding residue and affinity prediction.

MaskMol improves molecular image representation for activity cliff detection and compound potency prediction, outperforming existing methods across 20 targets.

This paper introduces GENiPPI, a molecular generative framework for PPI interfaces, which generates novel compounds targeting PPIs and demonstrates its effectiveness in capturing interface-molecule relationships.

GenoM7GNet is a deep learning model for predicting N7-methylguanosine sites using a nucleotide language model, achieving high accuracy and outperforming existing methods.

This study uses a transformer-based generative model to generate novel protein-protein complex conformations from molecular dynamics simulations, enhancing conformational exploration.

DeChat corrects errors in Nanopore R10 sequencing reads with high accuracy, preserving repeats and haplotypes, and improves genome assembly.

Amalga improves protein backbone designability by using folding and inverse folding guidance to generate more designable structures.

This paper introduces EADTN, an ensemble model for drug-target interactions that improves predictive performance and identifies false-negative samples through adaptive feature weighting and clustering fine-tuning.

Informeasure quantifies nonlinear dependence in biological networks using information theory measures, providing an R/Bioconductor package with multiple estimators for bivariate and trivariate analysis.

The paper quantifies synergistic and competitive miRNA interactions using multivariate information measurement, finding that neighboring and overlapping binding sites on mRNA lead to synergistic and competitive effects, respectively.

The paper introduces iEnhancer-CLA, a deep learning model for enhancer identification and strength prediction, showing improved accuracy and interpretability.

This paper introduces a structural perturbation method (SPM) to predict miRNA-disease associations, showing improved accuracy over existing methods in cross-validation and case studies.