Xin Gao

ReFeaFi predicts regulatory elements genome-wide using a two-model approach, achieving good performance across species and identifying disease-associated variants in human regulatory regions.

ChemPLAN-Net uses deep learning to predict novel inhibitor fragments for proteins by analyzing physicochemical features and co-crystal structures.

DeeReCT-TSS improves TSS annotation across cell types using DNA sequences and RNA-seq data, outperforming sequence-only methods and enabling cell-type-specific TSS identification.

DECCODE identifies small molecules that enhance cell reprogramming and conversion using transcriptional data and drug profiles.

The paper identifies related 2D chemical fingerprints that reduce molecular similarity scores and suggests their exclusion for improved virtual screening.

DeeReCT-APA quantitatively predicts alternative polyadenylation site usage across genes using a deep learning model, outperforming existing methods in multiple tasks.

This paper introduces NERO, a biomedical named-entity ontology with an annotated corpus, and demonstrates its utility through text embedding for discovering associations.

The paper develops a method to quantify rare genetic variants in human cells using long-read sequencing and shows CRISPR-Cas9 induces nonrandom large deletions and insertions.

This study uses genome-scale metabolic models and co-culture experiments to show that V. cholerae grows better and is more virulent during co-infection with ETEC due to metabolic cross-feeding.

The paper constructs a mathematical model of the IL-2-Teff-Treg system in SLE patients to simulate disease progression and identify therapeutic strategies for IL-2 dosing.

This paper reviews methods that integrate biomedical ontologies with machine learning to improve similarity computation and model performance using ontology embeddings and constraints.

The paper introduces a method that uses deep Siamese neural networks to normalize biomedical ontologies and improve predictions of protein interactions and gene-disease associations.

The paper evaluates how formal axioms and ontology metadata improve ontology-based predictions of protein-protein interactions and gene-disease associations.

The paper introduces iDTPnd, a computational method to identify novel drug targets by analyzing structural signatures of drug binding sites, validated with in vitro experiments on kinase inhibitors.

PointSite identifies protein ligand binding atoms using point cloud segmentation, outperforming previous methods and reducing false positives through cascaded filtering.

This paper introduces two novel algorithms for efficiently searching and mapping genomic subsequences in nanopore raw signals, improving SNP detection and haplotype classification under low coverage and accuracy.

The paper identifies LD motifs in the human proteome, reveals their role in cell adhesion and migration, and traces their evolutionary origin to a common ancestor.

The paper introduces omega, an RNA-seq quantification method that improves detection of transcriptional aberrations and reduces disease-gene search space using Mendelian disease cases.

PureseqTM predicts transmembrane topology from amino acid sequences, outperforming existing methods and correcting UniProt annotations in the human proteome.

This paper reviews deep learning applications in bioinformatics, covering methods, examples, and challenges in the big data era.

PGCN jointly trains graph convolutional network embeddings and association prediction on a heterogeneous gene-disease network to prioritize candidate disease genes and recover missing associations in OMIM.

The paper introduces cwDTW, a wavelet-based DTW algorithm for nanopore sequencing signal alignment, achieving high accuracy and speed compared to existing methods.

SupportNet addresses catastrophic forgetting in incremental learning by combining deep learning with SVM to retain old knowledge while learning new data, outperforming existing methods in multiple tasks.

DeepSimulator generates realistic Nanopore sequencing data using a context-dependent deep learning model, improving signal accuracy and enabling tool development.