Multi-Omics Target Discovery after Data Integration: An Umbrella Review of Causal Evidence, Reproducibility, and Translational Readiness Download PDF

Multi-omics integration is increasingly used to prioritize therapeutic targets by combining genomic, transcriptomic, proteomic, metabolomic, and related molecular evidence. Yet the accumulation of data layers does not necessarily produce stronger target claims. Reviews differ substantially in their definitions of integration, evidentiary standards, causal assumptions, validation requirements, and interpretation of translational readiness. This umbrella review evaluates how review-level evidence supports multi-omics target discovery and where conclusions remain constrained by methodological heterogeneity, overlapping primary studies, population underrepresentation, technical variability, and insufficient experimental confirmation. Eligible reviews and supporting methodological sources were organized through a review-of-reviews framework that distinguishes substantive evidence synthesis from reporting guidance and selected translational anchors. Review quality, search transparency, primary-study overlap, consistency, modality-specific contributions, causal inference, perturbational validation, reproducibility, diversity, druggability, and safety were treated as separate but connected dimensions. The synthesis indicates that genomic evidence can strengthen causal prioritization under explicit assumptions; transcriptomic and single-cell evidence can localize disease-relevant cellular contexts; proteomic evidence can support mechanism and target engagement; and metabolomic evidence can reveal phenotype-proximal pathway consequences. Integration becomes scientifically persuasive only when disagreement among modalities remains visible and when causal, perturbational, and cross-cohort validation are added. The principal contribution is a proposed target-evidence pyramid that separates association, contextual localization, causal support, experimental confirmation, reproducibility, and conditional translational readiness. The framework is an organizing synthesis rather than a validated scoring system. Its interpretation remains limited by incomplete review reporting, heterogeneous evidence structures, shared primary studies, and uneven availability of independent validation. Multi-omics integration should therefore be treated as a mechanism for evidence coordination, not as an automatic indication of pharmaceutical readiness.