Huiying Zhao

CellFM is a large-scale foundation model pre-trained on 100 million human cells, outperforming existing models in cell annotation, perturbation prediction, and gene function prediction.

The paper identifies genetic associations between diabetes medications and ECG features, finding a causal link between medication use and reduced T-wave area in lead aVR, increasing risk of ischemic heart disease and coronary atherosclerosis.

This paper introduces a graph convolutional network to predict disease-gene associations by leveraging external collaborative graphs and reducing noise through mutual information maximization.

GLMSite identifies DNA and RNA-binding sites using geometric graph learning on ESMFold-predicted structures, outperforming sequence-based methods and matching structure-based approaches.

The paper introduces RNALID, a database of 358 RNA-ligand interactions validated by low-throughput experiments, showing novel ligands and structural patterns with implications for drug discovery.

DGAT-cancer integrates mutation spectra and gene expression networks to identify 571 novel cancer drivers, with EEF1A1 validated as a glioma driver.

SPROF-GO predicts protein functions from sequences using a pretrained language model and label diffusion, outperforming existing methods by over 10% in AUPR.

LMetalSite predicts metal ion-binding sites from protein sequences using a pretrained language model and multi-task learning, achieving higher accuracy than structure-based methods.

This paper introduces UISNet, a deep learning framework for breast cancer outcomes prediction that improves accuracy and interpretability by integrating uncertainty-based gradients and biological knowledge.

TransEPI improves enhancer-promoter interaction prediction by using a Transformer model to capture long-range genomic dependencies across different cell types and chromosomes.

The paper introduces GraphSite, a method using AlphaFold2-predicted structures and graph transformers to predict DNA-binding residues, outperforming existing sequence and structure-based methods.

DGAT-onco detects oncogenes by integrating differential mutational analysis and functional impacts of somatic mutations, identifying 22.8% of its results as verified oncogenes.

This study uses Mendelian randomization to show a causal link between type 2 diabetes and cataract risk in East Asians, identifying two candidate genes involved in this relationship.

scAdapt improves single-cell RNA-seq classification across platforms and species by aligning domain distributions using adversarial learning.

The paper develops a deep learning framework to integrate multi-omics data for cancer prognosis, showing improved risk prediction across 15 cancers and identifying prognostic markers in breast cancer.

This paper introduces GraphSCI, a method that imputes dropout events in scRNA-seq data using graph convolution networks, outperforming existing methods on simulated and real data.

The paper develops a structure-aware model using graph convolutional networks and predicted contact maps to predict protein solubility from sequence, showing improved performance over existing methods.

The paper introduces TDimpute, a transfer learning-based deep neural network for imputing missing RNA-seq data using DNA methylation, showing improved accuracy across 16 cancer datasets.

This paper introduces TH-GRASP, an XGBoost-based method for predicting genome-wide RNA secondary structure profiles, achieving high AUC values and demonstrating effectiveness in reducing experimental noise.

The paper introduces MTDsite, a multi-task deep learning method for predicting protein-binding residues across DNA, RNA, peptide, and carbohydrate, achieving higher accuracy than existing methods.

This paper introduces SPROF, a method using 2D pairwise residue distance maps and image captioning to predict protein sequence profiles, achieving a 5.2% improvement over SPIN2.