Genetics

PRS and composite risk scores stratify cancer risk but have modest discrimination

Across cancer sites, polygenic risk scores (PRS) and composite risk scores (CRS) consistently stratify relative risk with monotonic increases across percentiles. Gains in discrimination (AUC/C-index) are generally modest and calibration varies substantially across populations. External validation and multi-ancestry evaluations are limited, raising concerns about transportability and equity. CRS are best used for risk stratification, targeted screening, prevention trials, and research designs rather than sole clinical decision tools. 1

Large biobank meta-analysis identifies 58 recessive gene–trait associations

In up to 948,690 sequenced individuals, statistical phasing increased detected bi-allelic damaging genotypes by 19% and found 5,563 genes with putative knockouts. Gene-based recessive testing identified 58 significant associations, with 17 showing stronger recessive than additive effects. Notable findings include recessive associations of HBB with heart failure and LECT2 with height. These results support routine consideration of bi-allelic coding variation in large-cohort genetic analyses and phenotype interpretation. 2

Genotype-guided SSRI prescribing increased 6‑month remission but not 3‑month response

In a pragmatic randomized trial of 1,460 patients, 692 had actionable pharmacogenetic phenotypes and were randomized to genotype-guided care or usual care. At 3 months there was no difference in PROMIS depression scores, adverse effects, or PHQ-8 change between groups. At 6 months, genotype-guided prescribing produced higher remission rates compared with usual care. Pharmacogenetic guidance may affect longer-term remission but not early symptom change; findings warrant replication and careful implementation. 3