The importance of intrafamilial cognitive phenotyping by the case of 22q11.2 deletion, 15q11.2 deletion, and families with inherited copy number variants of unknown significance

Background Rare copy number variants (CNVs) are known to be important contributors to the genetic cause of developmental differences (DD). Parental cognitive phenotyping could assist with interpreting inherited variants of uncertain significance (VUS), and could help to assess phenotypic variability and improve genetic counseling. However, no methodological framework exists for the intergenerational correlation of cognitive abilities. We introduce an approach to assess intrafamilial concordance of cognitive abilities and apply it to three trio cohorts: adults with de novo 22q11.2 microdeletion (n = 50) acting as a high penetrance control for the method, probands with inherited 15q11.2 BP1-BP2 deletions (n = 10) which is associated with a low penetrance of DD, and probands with DD and a rare CNV classified as VUS (CNVUS) and inherited from a parent with seemingly typical development (n = 21). Methods For each cohort, cognitive phenotyping of probands and both parents was performed using standardized, validated, and age-appropriate IQ tests (Wechsler scales). Siblings were tested when available. Intrafamilial concordance of cognitive abilities was assessed based on an overlap of 95% confidence intervals for full-scale IQ (FSIQ) and cognitive subdomains. Trio whole genome sequencing with variant analysis of known DD-related genes was performed in the CNVUS cohort to identify additional (likely) pathogenic variants associated with DD. Results For the de novo 22q11.2 microdeletion cohort, there was FSIQ discordance in most (42/50; 84%) proband-parent trios, a greater intrafamilial difference between proband and biparental FSIQ for probands with lower FSIQ (r = -0.684, p < 0.001), and evidence that the deletion has a more pronounced effect on performance than on verbal domains. In families with an inherited 15q11.2 deletion, there was evidence of assortative mating (proband's FSIQ aligned exclusively with that of the carrier parent in only two families), and intrafamilial variable expression. In the CNVUS cohort, 10 of 21 parents assessed were found to have borderline to mild ID, although considered within typical range at inclusion, and only five families (24%) showed concordance between proband and transmitting parent FSIQ. Reclassification of CNVUS, limited by small family size and assortative mating, was possible for two families to a likely pathogenic CNV, and for seven families to a likely benign CNV. WGS identified pathogenic variants contributing to the DD in two probands. Conclusion The results suggest that our approach for determining intrafamilial IQ correlation effectively captured the impact of de novo 22q11.2 microdeletion and additive parental background effect on cognitive impairment, consistent with the modest but detectable effect of 15q11.2 deletions. Parental cognitive data could assist with classifying inherited CNVs of unknown significance.