Diagnostic Yield of Genome Sequencing in an Iranian Exome-Negative Autosomal-Recessive Intellectual Disability Cohort

Intellectual disability (ID) affects approximately 1%-3% of the population and spans diverse clinical presentations with marked genetic heterogeneity, especially in consanguineous populations where autosomal-recessive ID is common. Despite advances in diagnostic methods, ~50% of individuals with ID remain without a molecular diagnosis. Genome sequencing (GS) can detect variant classes poorly captured by other methodologies, including deep intronic splice changes and structural variants. We performed short-read GS on 38 Iranian autosomal recessive intellectual disability (ARID) families that remained unsolved after exome sequencing (ES) and two phases of reanalysis. Sequencing and variant calling were performed on DRAGEN Bio-IT Platform with GRCh38 and variants were annotated in Golden helix Varseq software and the AnnotSV tool. GS yielded diagnoses in 2 of 38 families (5.3%), identifying a homozygous deep-intronic ATP8A2 variant (c.1473 + 519C > T) that creates a cryptic donor site and a 57-bp pseudoexon, and a homozygous ~103-kb CNTNAP2 deletion removing the in-frame Exon 2. Additionally, GS uncovered a homozygous NCOR1 missense variant that represents a novel candidate gene for ARID, but further studies are required to confirm the gene-disease association. The ATP8A2 and CNTNAP2 variants were uniquely detectable by GS. In conclusion, in a challenging, ES-negative ARID cohort, GS provided an additional ~5.3% diagnostic yield by uncovering a noncoding splice alteration and a large intragenic deletion. These results underscore GS as a valuable, though still modest, tool over ES and highlight significant interpretation challenges in noncoding regions. Continued advances in functional assays and complementary long-read technologies will be essential to further reduce the diagnostic gap in unresolved ID.