Chronic diseases impose a global health burden, contributing to high mortality and economic costs. Even with the recent surge in molecular data, these conditions remain largely incurable due to their biological complexity, data fragmentation, and analysis challenges, hindering early diagnosis, mechanistic understanding, and therapy. To address this, we developed PathoAnalyzer-I, an in silico platform that combines bioinformatics and machine learning to decipher chronic diseases. The tool requires no programming skills and provides a user-friendly interface for pathological analysis within a single framework. PathoAnalyzer-I uses a dataset of molecular data for 531 chronic diseases, from databases including the GWAS Catalog, PubChem, and STRING-db, enriched with machine learning–based predictions. Its dual-prediction system enhances molecular insights: one model imputes missing risk alleles with 77.6% accuracy, while a second predicts novel SNP–disease associations with 89.3% accuracy, providing avenues for future research. Applied to Alzheimer’s disease, the platform identified diagnostic biomarkers (e.g. rs6733839T), core genes in disease mechanisms (e.g. BIN1 , APOE , SLC24A4 , ABCA7 , PTK2B ), major pathological mechanisms like amyloid processing, synaptic dysfunction, and cellular vulnerability, as well as therapeutic molecules including Beta-Lapachone and preventive compounds such as curcumin. With its features, PathoAnalyzer-I enables scientists, regardless of their resources, to conduct in-depth in silico studies of chronic diseases.
Aguerd et al. (2026) studied this question.