Qinghua Cui

ASTRA enables zero-shot super-resolution of spatial transcriptomics across different capture geometries, improving accuracy and efficiency compared to existing methods.

The paper introduces i-Fold, a modified AlphaFold2 model that uses residue importance scores from ESM to improve protein structure prediction accuracy.

The paper develops a sperm-based bioassay using motility parameters to diagnose cardiovascular and cerebrovascular diseases with high accuracy.

The paper evaluates somatic contamination in germline genomes using WES and WGS data from UK Biobank, TCGA, and 1000 Genomes Project, revealing limitations in current filtering methods and the impact of somatic mutations on genetic studies.

This paper introduces BulkFormer, a foundation model for bulk transcriptome analysis that outperforms existing models in six downstream tasks.

CHIEF improves protein design by combining five models, achieving higher sequence recovery and lower perplexity, and successfully designing functional malate dehydrogenase.

The paper introduces MedImg, a database integrating 1,622,956 medical images from 103 datasets across 14 modalities and 9 organs, to support deep learning research in medical image analysis.

The paper introduces an improved Ka/Ks ratio (iKa/Ks) that accounts for the non-neutral effects of synonymous mutations on miRNA regulation, showing better performance in detecting selection pressures than the conventional Ka/Ks ratio.

The paper introduces PIC, a deep learning model for predicting human protein essentiality, outperforming existing methods and quantifying essentiality across humans, cell lines, and mice orthologues.

The paper introduces BIC, a tool to measure single base importance in human mRNAs and lncRNAs, showing its utility in pathogenicity assessment, cancer prognosis, and SARS-CoV-2 transmissibility prediction.

The paper investigates the functionality of synonymous mutations across species, finding that higher animal complexity correlates with increased functional impact through miRNA regulation.

This paper introduces a computational framework, GSGP, to decipher somatic mutational signatures and identifies novel aetiologies, including an RNA-derived signature in colorectal cancer.

The paper models a microRNA-derived disease network to identify methotrexate as a potential therapy for abdominal aortic aneurysm in mice.

The paper introduces WEAT, a weighted gene set enrichment tool that incorporates gene essentiality scores to improve functional analysis of gene lists.

The paper introduces a new metric, normalized bond energy (NBE), to quantify the druggable potential of chemical small molecules and validates it using DrugBank and Human Metabolome Database data.

The paper introduces sTAM, a tool for miRNA-set level disease biomarker discovery, showing its effectiveness in distinguishing cancers and predicting cerebrovascular disorders.

RPRH infers potential roles of plant miRNAs in humans and profiles their regulatory functions using a database of 10414 miRNAs and 18062 annotations.

DrugPattern is a web-based tool that performs drug set enrichment analysis to identify patterns in drug-related datasets, revealing unexpected protective roles of oxLDL in diseases.

GIC predicts the essentiality of protein-coding genes and lncRNAs using RNA sequence data, achieving high performance for lncRNAs in a mouse dataset.