Integrative Pathway Analysis in Complex Diseases

Abstract

Complex diseases including diabetes, cardiovascular disease, and neurodegeneration arise from the interplay of multiple genes, pathways, and environmental factors, and integrative pathway analysis that combines genetic association data with curated biological pathway knowledge can reveal disease mechanisms that single-gene analyses miss, yet the translation of pathway-level findings into actionable biological insight varies across analytical approaches. We evaluated 216 integrative pathway analysis programmes across centres in Austria, Estonia, and Germany between 2017 and 2021, spanning five method categories: over-representation analysis of gene sets, gene set enrichment analysis with ranked lists, topology-based pathway scoring, network propagation from seed genes, and multi-omics pathway integration. A Pathway Analysis Effectiveness Index (PAEI) was constructed from five sub-scores -- statistical power for pathway detection, biological specificity of findings, cross-study reproducibility, mechanistic interpretability, and therapeutic target nomination -- with weights from regression against sustained adoption into disease genomics pipelines. PAEI correlated with adoption at r = +0.84 and discriminated adopted from non-adopted methods with an AUC of 0.882. Gene set enrichment analysis scored highest (mean PAEI 0.824), while multi-omics integration trailed at 0.598. Only 35.2 percent exceeded the 0.75 threshold. Statistical power carried the largest weight (beta = +0.278), followed by cross-study reproducibility (beta = +0.230).