The Integration of Polygenic Risk Scoring in Forensic Psychiatry: Predictive Validity, Evidentiary Standards for Algorithmic Genomic Evidence, and a Proposed Auditability Framework

Polygenic risk scores (PRS) are moving from research settings toward clinical and forensic use. This review synthesises their predictive validity across the psychiatric domains most often at issue in forensic psychiatry: schizophrenia and the psychosis spectrum, bipolar disorder, major depressive disorder, ADHD and autism spectrum disorder, and borderline personality disorder (BPD). Accuracy varies sharply: schizophrenia PRS are among the strongest genomic predictors in psychiatry, yet specificity at flagged thresholds is barely above chance, and a landmark BPD genome-wide association study found SNP heritability of 17.3% while derived scores explained only 4.6% of phenotypic variance. The review maps the evidentiary environment for algorithmic evidence (Daubert, Frye, Federal Rules of Evidence 401-403 and 702, proposed Federal Rule of Evidence 707, and New South Wales Practice Note SC Gen 23) and proposes a three-part auditability standard for any genomic pipeline offered in court: cryptographic provenance, an algorithmic transparency log, and expert-in-the-loop certification. Augmentation, not replacement, is the defensible operating model. A conservative scenario suggests interception of roughly one-tenth of Australian suicide deaths if the undiagnosed fraction of attempters becomes visible to genomic screening.