Genetics

Rare and common variants converge at anthropometric loci, but ultra-rare variants add little heritability

Whole-genome sequencing of 672,976 individuals assessed rare noncoding, rare coding, and common variants for height, BMI, and waist–hip ratio. Ninety rare and low-frequency single-variant associations were identified, including two upstream IGF2BP2 variants that markedly lower waist–hip ratio with distinct adiposity profiles. Variant aggregation uncovered 135 coding and 51 noncoding signals, including FGF18 5′UTR variants with up to 6 cm height effects and UBR3 truncating variants increasing BMI by 2.7 kg/m². Ninety-seven percent of rare-variant associations localized near GWAS loci, and ultra-rare variants contributed only a small fraction of trait heritability. Heritability from these variants was largely shared across ancestries and concentrated around common-variant loci, supporting GWAS-based polygenic risk as capturing most genetic signal. 1

IBD colon eQTLs refine target genes at over 100 IBD loci

An eQTL study of colon tissue from 252 IBD patients mapped regulatory variants directly in diseased tissue. Integrating IBD and non-IBD colon eQTLs implicated 194 potential target genes across 108 IBD GWAS loci. IBD colon eQTLs were enriched for colocalization with IBD GWAS signals and nominated genes such as ABO and TNFRSF14 as candidate effectors. Disease-state eQTLs identified additional target genes beyond non-IBD tissue maps and showed larger effect sizes for some variants. These data support performing eQTL mapping in affected tissue to prioritize therapeutic targets and interpret IBD risk variants more accurately. 2

Semi-automated genomic newborn screening is feasible but challenged by penetrance and phenotype complexity

The NewbornsInSA study validated a genomic newborn screening workflow using whole-genome sequencing constrained to a 613-gene virtual panel. Retrospective samples with known variants confirmed technical performance, and bioinformatic scripts auto-classified cases lacking findings versus those needing manual review. Prospective implementation has already produced five reportable findings, demonstrating practical deliverability in a real-world cohort. Reporting dilemmas arose for genes causing multiple conditions, incomplete penetrance, or only mild phenotypes, complicating counseling and follow-up planning. These findings underscore that scalable genomic NBS depends as much on curated gene–phenotype frameworks as on sequencing and automation. 3

Rare coding variants in PTEN, DISP1, and SCUBE2 shape brain volumes and link to disease

Rare-variant gene aggregation in 50,061 individuals evaluated loss-of-function and missense variants across total and 43 regional brain volume phenotypes. Mutations in DISP1 and SCUBE2 associated with reduced cerebellar volume, likely via modulation of sonic hedgehog signaling. PTEN mutations associated with macrocephaly, consistent with PI3K–mTOR pathway dysregulation, and FA2H variants were hypothesized to influence cerebral white-matter volume. Seven genes showing population-level volume associations also overlapped rare brain disease genes in ClinVar, connecting subclinical variation with Mendelian phenotypes. These results support using rare-variant aggregation of imaging traits to refine disease genes and clarify neurodevelopmental mechanisms. 4