Rare SNVs identified with amino acid changes and computational predictions of pathogenicity.

1)PolyPhen2 scores close to 1 are likely to be pathogenic (highlighted in bold). HumDiv-trained Polyphen-2 assumes even mildly deleterious alleles as damaging to evaluate rare alleles potentially involved in complex phenotypes. 2)SIFT scores less than 0.05 are likely to be pathogenic (highlighted in bold). 3)GERP scores above 5 are highly conserved (highlighted in bold). 4)The Human Gene Mutation Database (HGMD) was searched to identify SNVs registered as disease causing mutations. Carrier frequencies of each SNV were statistically compared between ASD patients (cases) and in-house normal 212 controls (controls). Results are presented as odds ratio’s (OR) and p values. Pathogenic findings are shown in bold. CI, confidence interval; wt, wild type allele; mut, mutant allele; n.d., not determined.