Explainable Biomarker Discovery from RNA-Seq Data

Gradient-boosted trees (XGBoost) with exact TreeSHAP explanations find the genes in whole-blood RNA-seq that separate active tuberculosis (TB) from latent infection and healthy contacts. The study checks how stable the explained gene list is over 200 bootstrap models, builds a small gene panel from the most stable genes, and tests everything on four independent NCBI GEO cohorts from the UK, South Africa and India. A web application scores a new blood profile and shows why, gene by gene. A complete stand-alone MATLAB R2026a version for feature ranking and visualisation is included.

Category: Bioinformatics - Transcriptomics and Gene-Expression Analysis (Machine Learning, XGBoost, Explainable AI / SHAP, MATLAB, Python)

Transcriptomics & Gene-Expression Analysis Project B12 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.