Reference calibration, autozygosity and blood-derived variant interpretation in older adults

Population exomes contain signals of demographic history, inherited disease alleles and, when derived from blood, possible acquired variation. Distinguishing these signals is important in underrepresented populations, where reference absence can reflect both sampling and local frequency differences. To the best of our knowledge, this study represents the first Levantine old-age cohort characterized by whole exomes. We reanalyzed blood-derived exomes from 199 Lebanese trial participants aged 65-91 years, including 187 unrelated participants for frequency and recessive analyses. Twenty-nine participants (14.6%) reached the autozygosity expected for offspring of first cousins. Although 13.8% of observed autosomal alleles were absent at callable gnomAD Middle Eastern reference sites, size matching gave a reference-to-cohort single-nucleotide variant absence ratio of 0.98. Excess absence was concentrated among recurrent alleles; 360 were absent from the full reference, none classified as pathogenic or likely pathogenic in ClinVar. Separately, regional disease alleles showed nominal frequency differences, including MEFV p.Val726Ala (0.051 versus 0.012). Qualifying recessive alleles occurred in 55.1% of unrelated participants, decreasing to 45.5% after severity-related exclusions. Conditional qualifying biallelic genotype estimates increased from 2.0 to 27.7 per 1,000 offspring between unrelated-parent and first-cousin inbreeding coefficients. Twelve participants had clonal-driver candidates, but only five met both binomial and empirical allele-fraction criteria for the probable tier; a TP53 frameshift remained unresolved. Eight participants (4.0%) had unconfirmed, condition-matched candidate secondary findings. These observations distinguish reference sampling from allele-specific frequency differences and genotype expectations from clinical outcomes. They support complementary checks on population exome interpretation, while leaving generalizability, variant origin and clinical reportability to independent evaluation. Author SummaryWe studied genetic variation in 199 older adults from Lebanon to understand how population differences and aging can complicate the interpretation of genetic tests. A change missing from a reference database might be locally common, but it might also be missing simply because the database sampled too few people. We compared equally sized samples and found that the overall absence of observed changes was similar in the regional reference and another part of our study group. Individual disease-associated changes could nevertheless have different local frequencies. We also examined stretches of DNA inherited from shared ancestors and estimated how parental relatedness could change the chance of pairing disease-associated copies. These estimates describe genetic combinations, not predicted numbers of affected children. Finally, because blood cells can acquire changes during aging, we compared candidate changes with the sequencing patterns of other variants in the same data. Some findings remained uncertain, and none was clinically confirmed. Our results show why reference sampling, shared ancestry and possible acquired changes need separate consideration before drawing conclusions about inherited risk.