tools from the lab

MRanalysis - A Comprehensive Online Platform for Integrated, Multi-Method Mendelian Randomization and Associated Post-GWAS Analyses

posted on October 22, 2024

by Abao Xing. Introduction Mendelian randomization (MR) has emerged as a powerful epidemiological method for inferring causal relationships between exposures and outcomes using genome-wide association study (GWAS) summary data. By...

GWASkit - a tool designed to provide fast all-in-one preprocessing for GWAS summary data files.

posted on October 1, 2024

by Abao Xing. Introduction MRanalysis is freely available at https://mranalysis.cn. GWASkit is freely available at https://github.com/Li-OmicsLab-MPU/GWASkit. Mendelian randomization (MR) has emerged as a powerful epidemiological method for inferring causal relationships...

MIGOE

MIGOE: Minimum Information for Generative Omics Evaluation

posted on March 2026

A framework establishing rigorous standards for evaluating generative AI models in omics research. MIGOE defines six core principles (Tier Alignment, Data Provenance, Baseline & Shortcut Audit, Uncertainty Quantification, Downstream Biological Validity, Open Science) to ensure trustworthy and auditable clinical AI deployment.

THP-CNN

THP-CNN: Two-Stage Hierarchical Pruning of Convolutional Neural Networks for Rapid Pathogenic Bacterial Detection

posted on September 16, 2025

A novel two‑stage hierarchical pruning method for CNNs, enabling rapid and accurate detection of pathogenic bacteria from high‑resolution colony images in intensive care units.

ClinMetML

ClinMetML: End‑to‑End Automated Framework for Metabolomics Biomarker Discovery

posted on March 3, 2026

A flexible Python framework for metabolomics that integrates rigorous data cleaning, multi‑stage feature selection, multicollinearity reduction, model training (including deep learning), SHAP interpretability, and Decision Curve Analysis (DCA). Streamlines robust clinical predictive modeling.