Yu Zhao

Alignment-Aware Genomic Coordinate Encoding for Long-Range DNA Modeling

This paper introduces USM-AA, an alignment-aware genomic foundation model that improves long-range DNA modeling by incorporating reference genome coordinates and reverse-complement masked learning.

Early and late RNA eQTL are driven by different genetic mechanisms

The paper identifies distinct genetic mechanisms for early (nuclear) and late (cellular) RNA eQTL by comparing their genomic localization and functional effects across tissues.

Cancer-like fragmentomic characteristics of somatic variants in cell-free DNA

This study identifies cancer-like fragmentomic traits in somatic variants of cfDNA, enabling cancer diagnosis through integrative genomic and fragmentomic analysis using AI models.

A population-scale atlas of blood and tissue in lupus nephritis

A single-cell atlas of blood and tissue samples from lupus nephritis patients and healthy controls identifies stromal and immune populations, including scar-associated macrophages, which correlate with disease activity.

Candida albicans drives colorectal cancer progression by inducing hypoxia signaling

This study shows Candida albicans promotes colorectal cancer by activating hypoxia signaling through EGFR/TLR2 pathways.

An iPSC model of fragile X syndrome reflects clinical phenotypes and reveals m6A-mediated epi-transcriptomic dysregulation underlying synaptic dysfunction

This study uses iPSC-derived neurons to show that FMRP deficiency in fragile X syndrome leads to m6A-mediated transcriptome dysregulation, causing synaptic dysfunction and suggesting therapeutic targets.

CD-GPT As a Biological Foundation Model Bridging the Gap between Molecular Sequences Through Central Dogma

CD-GPT is a foundation model that captures sequence relationships between DNA, RNA, and proteins, performing well in predictive and generative tasks across molecular types.

Multi-Level Attention Graph Neural Network for Clinically Interpretable Pathway-Level Biomarkers Discovery

The paper introduces MLA-GNN, a graph neural network that discovers clinically interpretable pathway-level biomarkers by integrating multi-level graph features and a saliency mechanism, achieving state-of-the-art results on transcriptomic and proteomic data.