Zhan Mei Zhou

ORIGAMI reconstructs cis-regulatory networks using multi-omics data and deep learning, revealing synergistic transcription control and disease-associated rewiring.

ProSiteHunter improves protein binding site prediction by integrating sequence-based features and a fine-tuned language model, achieving better performance than existing methods.

CHORD integrates cross-species single-cell transcriptomic data to learn gene, cell, and cell-type representations, enabling cell-type annotation and cross-species comparison of developmental and phenotypic patterns.

This paper introduces LucaOne, a foundation model trained on nucleic acid and protein sequences, demonstrating understanding of biological principles and performing well on related tasks.

TRAIT compiles nearly 8 million experimentally validated TCR-antigen pairs to provide a comprehensive resource for T-cell receptor interactions, aiding immunotherapy and vaccine design.

DeepCIP predicts circRNA IRESs using sequence and structure information, outperforming existing methods and improving predictive performance through feature integration.

DeepTAP predicts TAP-binding peptides using an RNN, outperforming existing methods on benchmark datasets.

Spanve is a nonparametric statistical method that identifies spatially variable genes in spatial transcriptomic data with fewer false positives, enabling better downstream analysis.

TSNAdb v2.0 updates the tumor-specific neoantigen database with stricter identification criteria, predicted neoantigens from three mutation types, and experimentally validated neoantigens classified by evidence.

MATTE identifies differentially expressed genes and modules in noisy transcriptome data, improving analysis of phenotype-related biological insights.

The paper predicts potential B/T-cell epitopes for multiple coronaviruses and stores the results in a database for vaccine development.