Di Wu

GARD: Genomic Data based Drug Repurposing in Head and Neck Cancer with Large Language Model Validation

The paper develops GARD, a pipeline using genomic data and large language models to identify drug repurposing candidates for head and neck cancer, finding both known and emerging therapeutic options.

Differential roles of Rad18 in repressing carcinogen- and oncogene-driven mutagenesis in vivo

This study shows Rad18 suppresses 4NQO-induced oral carcinogenesis but not KrasG12D-driven lung tumors, highlighting context-specific tumor-suppressive roles.

The RING Finger E3 Ligase RNF25 Protects DNA Replication Forks Independently of its Canonical Roles in Ubiquitin Signaling

This study identifies RNF25's role in protecting DNA replication forks from degradation, independent of its ubiquitin signaling functions.

CD38⁺ Endothelial Remodeling Defines Spatially Diverse Vasculopathy Programs in Rapidly Advancing Oral Inflammation

The paper identifies CD38⁺ endothelial remodeling as a spatially distinct vasculopathy feature in rapidly advancing oral inflammation using scRNA-seq and spatial proteomics.

High-dimensional Biomarker Identification for Scalable and Interpretable Disease Prediction via Machine Learning Models

This paper introduces HiFIT, a framework for high-dimensional biomarker identification that improves disease prediction and feature importance detection using machine learning.

Deep-learning-based interpolation of longitudinal microbiome data powers biologically informative discovery

This paper introduces BGOB, a deep learning framework for longitudinal microbiome interpolation, improving analysis of microbial community trajectories and revealing disease-related associations.

MAPK14/p38α Shapes the Molecular Landscape of Endometrial Cancer and promotes Tumorigenic Characteristics

This study identifies MAPK14/p38α as critical for HGEC spheroid growth, shaping its transcriptomic, proteomic, and metabolomic landscapes and promoting tumorigenic properties.

Longitudinal Microbiome-based Interpretable Machine Learning for Identification of Time-Varying Biomarkers in Early Prediction of Disease Outcomes

This paper introduces LP-Micro, a machine learning framework for analyzing longitudinal microbiome data to identify time-varying biomarkers predictive of disease outcomes.

Metabolic reprogramming of immune cells in HIV infection and treatment

This study examines metabolic reprogramming in immune cells during HIV infection and after antiretroviral therapy using single-cell RNA sequencing and pathway analysis.

Comprehensive evaluation of methods for differential expression analysis of metatranscriptomics data

The paper evaluates ten differential expression methods for metatranscriptomics data using real and simulated datasets, finding the logistic-beta test has highest sensitivity and reasonable type I error control.

A bivariate zero-inflated negative binomial model and its applications to biomedical settings

The paper develops a bivariate zero-inflated negative binomial model and applies it to analyze zero-inflated count data from scRNA-seq and dental caries studies.

BZINB model-based pathway analysis and module identification facilitates integration of microbiome and metabolome data

The paper introduces a BZINB model-based method for integrating microbiome and metabolome data, showing improved accuracy in identifying correlations compared to traditional methods.

Differential Expression Analysis for Metatranscriptomics with Sample-Paired Metagenomics Data

The paper introduces MetaDePair, a statistical model for differential expression analysis in metatranscriptomics using paired metagenomics data, improving accuracy and reducing false discoveries.

Jointly analyzing the association of human milk nutrients with cognition and temperament traits during the first 6 months of life

The paper examines how various human milk nutrients are associated with cognitive and temperament traits in infants up to 18 months old, finding persistent effects of certain nutrients on development.

TWO-SIGMA-G: A New Competitive Gene Set Testing Framework for scRNA-seq Data Accounting for Inter-Gene and Cell-Cell Correlation

TWO-SIGMA-G tests gene sets in scRNA-seq data by accounting for gene and cell correlations, showing improved power and error control in simulations and real datasets.

Dimensionality reduction by UMAP reinforces sample heterogeneity analysis in bulk transcriptomic data

This study compares dimensionality reduction methods for bulk transcriptomic data, finding UMAP better preserves sample neighbourhood and clustering accuracy than PCA, MDS, and t-SNE.

FIREcaller: Detecting Frequently Interacting Regions from Hi-C Data

FIREcaller detects frequently interacting regions from Hi-C data and shows their association with gene regulation and GWAS variants in human tissues.

TWO-SIGMA: a novel TWO-component SInGle cell Model-based Association method for single-cell RNA-seq data

TWO-SIGMA is a single-cell RNA-seq differential expression method that handles drop-out, overdispersion, and within-sample correlation, outperforming existing methods in simulations and real data.

limma powers differential expression analyses for RNA-sequencing and microarray studies

limma couples gene-wise linear models with empirical Bayes variance moderation and weights; voom extends it to RNA-seq, and added functions test differential splicing and gene sets.