Leyi Wei

The paper shows DNA sequence determines the destination of spontaneous epimutations in plants, using a deep-learning model trained on one methylome to predict methylation set points across species.

This paper introduces RA-GHN, a framework for antibody sequence-structure co-design that captures multi-body interactions and structural constraints, outperforming existing methods.

This paper introduces LARK, a framework that learns chemical representations from reactions, showing improved performance in reaction prediction and molecular property prediction.

GeoEPred improves effector protein prediction by integrating sequence and structural information to better model functional motifs and spatial conformations.

Melody predicts DNA methylation from genomic sequences across human tissues and outperforms existing methods, also generalizing to methylation quantitative trait locus effect prediction.

The paper introduces RiboSeek, a framework for RNA 3D structure search using geometric alphabets, achieving fast and accurate comparisons.

GeoARG improves detection of distant antibiotic resistance genes using structural features and protein language models, uncovering 1,485 novel candidates.

scMomer is a pretraining framework for single-cell multi-omics modeling under missing modality conditions, generating biologically meaningful embeddings and outperforming unimodal approaches in downstream tasks.

MultiNano uses multi-view deep learning to accurately detect and quantify m6A RNA modifications from Nanopore DRS signals, achieving state-of-the-art performance in multiple tasks.

DiffusionST improves spatial transcriptomics data quality and identifies spatial domains through a diffusion model and graph convolutional network.

HGCPep improves ncPEP identification by using a hypergraph model to capture peptide-RNA relationships, outperforming existing methods.

DenoiseST is a deep learning method for spatial transcriptomics that improves data imputation, clustering, and identification of functionally variable genes by reducing dropout noise.

StructuralDPPIV is a deep learning model that predicts DPP-IV inhibitory peptides using molecular graph and sequence features, outperforming existing methods on test datasets.

This paper introduces MucLiPred, a deep learning model using dual contrastive learning to improve prediction of protein-ligand interactions and identify binding residues, showing superior performance over existing methods.

DeepBIO is a deep-learning platform that automates biological sequence analysis, offering 42 algorithms and visualization tools for functional prediction and annotation.

This paper introduces PHAT, a deep graph learning framework for peptide secondary structure prediction, achieving better accuracy and interpretability for short peptides.

PRISMOID is a 3D structure database for post-translational modifications and mutations, providing structural and functional insights into PTM sites and their interactions.