Yungang Xu

DyMoTree infers early cell fate transitions and drivers from single-cell transcriptomes using a tree-structured neural network, outperforming existing methods in resolving fate biases.

CSFeatures identifies cell type-specific differential features in single-cell and spatial omics data by considering expression levels, smoothness, and proportion across cell populations.

This study uses single-cell RNA sequencing to show that chronic methamphetamine exposure alters liver immune cell composition and function, with involvement of the dopamine D1 receptor.

A GPT-2 retrained from scratch on protein-ligand binding data, with a BPE tokenizer redesigned for proteins and ligands, is proposed for generating ligands against a target protein.

scIGANs uses GANs to impute dropouts in scRNA-seq data, improving downstream analysis and performance across datasets.

The paper identifies histone modifications associated with alternative splicing events in human ESCs and shows their role in cell cycle regulation and fate decision.

The paper introduces DeepCode, a deep learning model that integrates epigenetic features to predict splicing patterns during hESC differentiation and identifies a novel mechanism linking histone modifications to cell fate decisions.