BRPtools: An AutoML-Powered web platform for multiclass disease prediction from bulk blood RNA-seq data

Blood-based transcriptomic profiling provides a minimally invasive approach for disease diagnosis; however, the integration of large-scale, heterogeneous RNA sequencing (RNA-seq) datasets remains challenging. Here, we manually curated and uniformly reprocessed 134 publicly available human RNA-seq datasets (n = 9,872 samples), covering 88 distinct blood-related diseases from whole blood and peripheral blood mononuclear cell data. To enable robust and scalable multiclass prediction, we employed AutoGluon, an ensemble-based AutoML framework that automates model selection and hyperparameter tuning through multilayer stacking. Our models achieved high accuracy across most disease categories (>90%), demonstrating strong generalizability. To support biological interpretation, we also performed differential expression and pathway enrichment analyses, revealing disease-associated signatures enriched in relevant Gene Ontology terms and Kyoto Encyclopedia of Genes and Genomes pathways. We implemented these capabilities in BRPtools, a web-based platform featuring two modules: (1) a search module for exploring gene- and disease-level expression profiles and (2) a predict module for disease classification using user-uploaded RNA-seq count matrices. The platform supports diagnostic inference and validation and is designed to be accessible to users without programming experience. BRPtools offers an integrated, interpretable, and user-friendly solution for transcriptome-based disease prediction, supporting applications in biomarker discovery, digital diagnostics, and precision medicine.