Variant filtering
STU-66 family: ES data of the four members, resulted in 73,426 variants in affected father, 73,689 in unaffected mother, 73,141 in unaffected sibling and 73,198 variants in proband.
Applying variant filter criteria (detailed in methodology), resulted in a mean of 2245 variants (ranging from 2188-2296). The variant data of the four individuals were analyzed using Venn diagram that produced 15 possible combinations of the data (figure 3). One such combination revealed 276 genes that were common to affected father and the proband (figure A1). On filtering further for the common variants in them ensued 260 variants harbored in 248 genes. Of these 260 common variants, 218 were previously known (dbSNP database) and 42 were novel variants (hitherto unreported). Only 85 of the 218 known variants and 12 of the 42 novel variants fall in the coding region. Applying a four step exclusion criteria (detailed in methodology) to known variants, shortened the list to 18 genes (all heterozygous variants). The genes of interest thus sums up to 19 known variants and 12 novel variants, spread across 30 genes (table A1).
STU-65 family: ES data of the two affected sibs (V-33, proband and V-35, affected brother) resulted in 58,346 and 51,062 variants respectively. Applying similar criteria as described above using Venn diagram approach, 294 genes were found common to the sib-pair (figure A2). The downstream analysis of the variants observed in the 294 common genes is depicted in the flow chart (figure A3). Those variants common to both sibs were sorted out resulting in 161 common variants harbored in 153 genes. Of these, 158 were known variants and 3 were novel variants that were all found in the coding regions. Applying the same four step exclusion criteria (as in STU-66) to 158 known variants, narrowed down the gene list to 34 variants (4 homozygous and 30 heterozygous). Thus the genes of interest sum up to 31 known and 3 novel variants, in 34 genes. (table A2).
No significant intronic/UTR variation was detected, other than the indels and missense variations, in the exome data of both families. Considering stuttering to be a less common trait, an allele count of more than 20 in gnomAD database were excluded, narrowing the list to 35 genes.