Abstract Tuberous sclerosis complex (TSC) is a genetic multisystem disorder regarded as having near-complete penetrance, a view largely derived from clinically ascertained cohorts. As genomic newborn screening is piloted internationally, robust estimates of penetrance and expressivity in unselected populations are urgently needed. We analysed genome sequencing data from >900,000 adults in UK Biobank (UKB) and the All of Us Research Program (AoU), representing multiple genetic ancestries, to identify individuals carrying high-confidence ACMG/AMP (likely) pathogenic TSC1 / TSC2 variants. Electronic health records and self-reported data were interrogated for TSC-related phenotypes, including diagnostic codes, and evidence of epilepsy, chronic kidney disease and other hamartomas. We identified 61 adults with (likely) pathogenic TSC1 / TSC2 variants and phenotype data available. Only 20/61 (33%) had a recorded diagnosis of TSC and a further 8/61 (13%) had at least one TSC phenotype; 33/61 (54%) had no recorded TSC diagnosis, manifestations or suggestive medication history, despite median age ≥55 years (UKB ≥ 75 years) with extensive longitudinal healthcare data. Apparent reduced penetrance was observed for TSC1 in both cohorts (46.6% UKB, 46.1% AoU), consistent with previously reported milder phenotypes associated with this gene. Apparent TSC2 penetrance was also incomplete (27.8% UKB, 66.7% AoU, p = 0.038, not significant after Bonferroni correction). Borderline significant apparent penetrance differences between UKB and AoU were not explained by gene, variant class, or ClinVar designation, but may reflect healthy volunteer bias in UKB. Overall, these findings challenge the assumption of near-complete penetrance for (likely) pathogenic TSC1 / TSC2 variants and have important implications for genomic newborn screening, variant interpretation and counselling of asymptomatic carriers.
Fasham et al. (2026) studied this question.