Experimental validation of AI-prioritized therapeutic targets: a systematic review and descriptive quantitative synthesis

He, Yuanzhi ORCID: https://orcid.org/0009-0007-8424-2654, Xu, Jiayu and Wang, Yuheng 2026. Experimental validation of AI-prioritized therapeutic targets: a systematic review and descriptive quantitative synthesis. International Journal of Drug Discovery and Pharmacology Item availability restricted.

	PDF - Accepted Post-Print Version Restricted to Repository staff only Download (775kB)
	PDF (Provisional file) - Accepted Post-Print Version Download (17kB)

Abstract

Artificial intelligence (AI) is used to prioritize therapeutic targets, but support for a selected target differs from evidence that AI improves selection. We synthesized peer-reviewed reports linking AI-informed molecular target selection to new disease-relevant functional testing. PubMed/MEDLINE and Scopus were searched to 27 July 2026, with citation tracking and supplemental discovery searches on 15 October 2026. Two human reviewers independently assessed eligibility in the original and supplemental search sets. The retained evidence comprised 305 reports grouped into 303 research families, including 283 additional families. Highest experimental settings were cell-based testing in 163 families, tissue-context models in 2, animal efficacy in 137 and prospective clinical evaluation in 1. Comparable candidate panels were available for 283 families; 265 tested a single candidate, 8 tested two and 10 tested at least three. We report panel-size and candidate-unit strata, equal-family summaries and sensitivity analyses, with experimental setting separate from six-domain appraisal. Selected panels frequently reported only supported candidates, while leakage, incomplete reporting and heterogeneous outcome definitions limited inference. Published evidence supports target-specific functional effects in defined systems. Candidate-level validation fractions summarize the reported experimental panels. Future evaluations require complete candidate accounting and matched prospective comparisons. The review was retrospectively registered with OSF (doi: 10.17605/OSF.IO/BHX6N).

Item Type:	Article
Status:	In Press
Schools:	Schools > Computational & Mathematical Sciences Schools > Computer Science & Informatics
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science R Medicine > RM Therapeutics. Pharmacology
Publisher:	Scilight Press
ISSN:	2653-6234
Related URLs:	Publisher
Date of First Compliant Deposit:	10 November 2026
Date of Acceptance:	23 October 2026
Last Modified:	10 Nov 2026 10:15
URI:	https://orca.cardiff.ac.uk/id/eprint/190751

Actions (repository staff only)

Edit Item