Population-Specific Carrier Frequencies in an Underrepresented Genetically Heterogeneous Population: Implications for Expanded Carrier Screening.

Background/Objectives: Expanded carrier screening (ECS) panels often rely on pan-ethnic databases, such as the ACMG Tier 3 panel, but their ability to capture population-specific reproductive risk is uncertain. We characterized pathogenic/likely pathogenic (P/LP) variants and at-risk-couple (ARC) probabilities in an underrepresented population from North Macedonia to inform ECS gene prioritization. Methods: NGS data from 1438 subjects of Macedonian, Albanian, and other local ancestry, referred for rare-disease diagnostics, were analyzed across 1795 candidate autosomal recessive/X-linked genes. P/LP variants were compared with gnomAD exomes v2.1.1 non-Finnish European frequencies (Fisher's exact test). Gene-specific ARC probabilities ranked genes and generated cumulative ARC curves per subgroup. Results: We identified 2007 P/LP variants across 1096 genes. One hundred variants in 89 genes reached an allele frequency ≥ 1/200 in the Macedonian or Albanian subgroups; over one-third were absent from candidate ECS panels, mostly linked to severe or variable phenotypes. ARC saturated rapidly: genes with carrier counts ≥ 1/100 (ACMG Tier 2) accounted for over 85% of total ARC, and with ≥1/200 (ACMG Tier 3) for more than 93%. Total ARC among autosomal recessive panel genes (X-linked genes analyzed separately) was 3.42%, rising to 5.02% in the Albanian subgroup. Excluding top hypomorphic variants reduced ARC by 34-37%, highlighting variant-specific penetrance. The highest-ARC genes diverged between subgroups, with limited overlap with pan-ethnic panels. Conclusions: Population-specific carrier frequencies influence gene prioritization and estimated reproductive-risk probabilities for ECS. Population-tailored ECS panels that incorporate local carrier-frequency data, including embryonic/fetal-lethal genes, and account for hypomorphic alleles to avoid overestimating reproductive risk probabilities could improve the identification of couples at increased reproductive risk, particularly in underrepresented populations.