Haixu Tang

This study identifies shared age-associated pathway remodeling across species using comparative metabolomics, highlighting lipid and nucleotide pathways as core features of aging.

The paper introduces TopRepo, a comprehensive TD-MS spectral repository, and shows it improves proteoform identification and supports DL model training for spectral prediction.

The paper benchmarks toxicity prediction models and introduces a 3D-structure-based deep learning model that improves classification accuracy.

FIDDLE is a deep learning method that predicts chemical formulas from tandem mass spectra, achieving high accuracy and speed compared to existing methods.

The paper develops a machine learning model to predict SARS-CoV-2 mutation trends using historical data, achieving low prediction error and robustness across mutation patterns.

This study introduces a DNA foundation model to assess functional genetic variation across diverse populations, demonstrating its utility in understanding Alzheimer's disease risk and ancestry-specific gene expression.

This paper introduces a task-specific transfer learning approach to improve small molecule retention time prediction with limited data, achieving higher accuracy across diverse experimental conditions.

TopLib is a software package for building and searching top-down mass spectral libraries, offering faster and more reproducible proteoform identification than conventional methods.

This paper introduces a panproteome graph to enhance peptide and protein identification in proteomics and metaproteomics by leveraging the human pangenome.

The paper introduces DCTdomain, a method that segments proteins into domains and uses discrete cosine transformation to create domain-level embeddings for improved protein similarity search.

The paper introduces a scalable algorithm for clonal reconstruction from sparse time course genomic sequencing data, achieving fast and accurate results on simulated and real data.

3DMolMS predicts tandem mass spectra from 3D molecular conformations, achieving high cosine similarity with experimental spectra and enabling generalized predictions across instruments.

MORONET is a multi-omics integration method using graph convolutional networks for biomedical classification, outperforming existing approaches and identifying disease-related biomarkers.

This study uses mutation accumulation and whole-genome sequencing to reveal new complexities in SOS-induced untargeted mutagenesis in Escherichia coli, showing sequence and DNA-strand bias in transversion mutations.

This paper develops algorithms to reconstruct clonal evolution from time course genomic sequencing data, achieving accurate results on both simulated and experimental data.

msCRUSH clusters tandem mass spectra using LSH, achieving faster speed and better sensitivity than existing tools, improving peptide identification.

This study uses whole-genome sequencing of MMR-defective E. coli to identify base-pair substitution patterns and their determinants, revealing context-dependent mutational hotspots and strand bias.

The paper introduces a computational framework integrating genomic, transcriptomic, and proteomic data to identify and interpret translated fusions and micro structural variations in cancer, finding many novel peptides and private aberrations in breast cancer.