Yixue Li

This paper introduces GeneEmbedLLM, a framework using large language models to generate gene embeddings from multi-source textual annotations, achieving competitive performance in various downstream tasks.

AutoBioKG constructs context-aware biomedical knowledge graphs to improve scientific inference by encoding environmental conditions and entity attributes, outperforming baselines in multiple benchmarks.

TSTScope integrates single-cell transcriptomics and T-cell receptor data to identify functional T cell states predictive of immunotherapy response, finding that functional state, not abundance, determines treatment outcomes.

DERIVE predicts antigenic change across viruses using disentangled latent representations of evolutionary history, showing strong transferability to diverse viral families.

This study introduces a method to predict clinical phenotypes from single-cell data using attention-based multiple instance learning, demonstrating its effectiveness in predicting disease severity and age-related phenotypes.

This study benchmarks deep learning methods for single-cell RNA sequencing integration, identifies limitations in existing metrics, and proposes improvements to better preserve biological signals.

scTFBridge infers gene regulatory networks from single-cell multi-omics data by integrating TF-motif binding knowledge, outperforming baselines in cis- and trans-regulation tasks.

ScReNI integrates scRNA-seq and scATAC-seq data to infer cell-specific regulatory networks and identify cell-enriched regulators.

EAP provides a cloud-based platform for analyzing large-scale ChIP/ATAC-seq data, generating publication-ready results and enabling epigenomic analysis in cancer and therapeutic target discovery.

BioOS provides a cloud-native platform for reproducible biomedical research by integrating data, code, tools, and environments into workspaces, enhancing data integrity and accessibility.

The paper introduces a framework that combines large language models with knowledge graphs to improve factual accuracy in biomedical question answering, demonstrating effectiveness in drug repositioning and resistance prediction.

This paper introduces a bioinformatics copilot system that combines AI and human expertise to enhance large-scale omics research and uncover biological insights.

This study introduces XgeneVAE, a deep learning framework that characterizes tumor heterogeneity and fitness effects using multi-omics data, showing its ability to reflect patient survival differences and uncover genomic alterations.

This study performs genome assembly and transcriptome analysis of Microtus fortis to understand its resistance to Schistosoma japonicum infection, revealing differences in immune response timing and effectiveness compared to mice.

DComboNet prioritizes effective anti-cancer drug combinations using a two-level network model, achieving high AUC performance and inferring potential combination mechanisms.

This paper models the aging self-organization system using supervised learning to optimize aging networks and study molecular mechanisms of aging and related diseases.

The paper identifies base editor variants with reduced off-target effects on both DNA and RNA.

This study sequenced 128 camels across Asia to trace the origin and migration of domestic Bactrian camels, finding an Iranian origin followed by eastward migration.

This study compares off-target single nucleotide variants (SNVs) from CRISPR-Cas9 and base editing (BE3) in mouse embryos using the GOTI method, finding BE3 causes significantly more SNVs.

This study identifies four pan-cancer evolutionary patterns linked to clinical aggressivity using Bayesian mutation landscape analysis across 21 cancer types.