Ting Chen

Escaping the drug-bias trap: using debiasing design to improve interpretability and generalization of drug-target interaction prediction

This paper introduces UdanDTI, a deep-learning model for drug-target interaction prediction that addresses the drug-bias trap, showing improved accuracy and interpretability across datasets and EGFR mutation subgroups.

Rag2Mol: Structure-based drug design based on Retrieval Augmented Generation

This paper introduces Rag2Mol methods for structure-based drug design using retrieval-augmented generation, producing drug candidates with better binding affinities and drug-likeness than existing models.

Extraction of biological signals by factorization enables the reliable analysis of single-cell transcriptomics

This paper introduces scPhenoXMBD, a deep auto-regressive factor model that extracts biological signals from single-cell transcriptomics data, improving analysis reliability and enabling better identification of IFNβ-responsive pathways.

scPheno: A deep generative model to integrate scRNA-seq with disease phenotypes and its application on prediction of COVID-19 pneumonia and severe assessment

scPheno integrates scRNA-seq with disease phenotypes to identify transcriptional variations linked to COVID-19 diagnosis and severity, achieving 81.3% accuracy using 4 genes.

Inferring Multiple Metagenomic Association Networks based on Variation of Environmental Factors

The paper introduces kLDM, a model that infers multiple metagenomic association networks under varying environmental conditions, showing improved performance over traditional methods.

redPATH: Reconstructing the Pseudo Development Time of Cell Lineages in Single-Cell RNA-Seq Data and Applications in Cancer

redPATH reconstructs pseudo developmental time of cell lineages in single-cell RNA-seq data and identifies stem cell-like malignant glioma cells and key regulatory genes.

Determine independent gut microbiota-diseases association by eliminating the effects of human lifestyle factors

The paper uses machine learning to assess gut microbiota-disease associations while accounting for human lifestyle factors, finding independent associations with IBD and unhealthy status but not others.

mLDM: a new hierarchical Bayesian statistical model for sparse microbioal association discovery

The paper introduces mLDM, a hierarchical Bayesian model to discover microbial associations in metagenomic data while addressing compositional bias and environmental factors.

ProbAlign: a re-alignment method for long sequencing reads

ProbAlign is a re-alignment tool for long sequencing reads that improves variant calling and viral quasispecies reconstruction by reducing false positives and increasing true positives.