16472
Whole Exome Sequencing of ASD in Korean Population
Objectives: The objective of this pilot study is to examine genetic variant of ASD using WES in family samples from Korean population and developing analytic pipeline for further studies with larger samples.
Methods: For ascertainment of subjects, we used Korean versions of Autism Diagnostic Observation Schedule (ADOS) (Lord et al., 2008) and Autism Diagnostic Interview-Revised (ADI-R) (Lord et al., 1994) along with Social Communication Questionnaire (SCQ) (Rutter et al.,2003) and Social Responsiveness Scale (SRS) (Constantino, 2005). We identified 13 families with ASD, composed of probands, biological parents and unaffected siblings, from pooled database of Korean Autism Genetic Study Consortium. We selected severe, typical subjects of ASD for attaining relative homogeneity of the phenotype, with criteria of; 1) probands meeting full criteria of autistic disorder of DSM-IV-TR, 2) age between 48~156 months, 3) nonverbal or phrase speech level of language development, and 4) scores of lifetime algorithms of ADI-R and SRS are within 25 percentile of the database.
We had performed whole-exome sequencing for 13 probands, minimum 50X on target. At least 94% of target area were covered more than 5 sequence reads. In order to find de novo mutation candidates, we had performed additional high-coverage whole-exome sequencing on two pooled samples from mothers (250 million paired-end reads) and fathers (236 million paired-end reads) of probands. All the sequence reads were mapped onto the human reference genome (hg19 without Y chromosome). Variant discovery had been done by BWA, Picard, GATK, and in-house custom annotation pipeline. We had selected de novo mutation candidates (nonsynonymous, splice site, and coding indel mutation) from each proband which are not detected in two pooled samples and not reported in dbSNP137 and in-house Korean databases.
Results: We had selected 29 de novo mutation candidates from 21 genes; ABCF3, ADRB1, AKNA, AKT1S1, CELSR3, CHKA, FOXK2, IFI27L2, ITFG3, KCTD9, KDM6B, LOXL2, MYH14, PAX2, POLRMT, RBM20, SLC47A2, SOX7, TCTE1, TXNDC11, and USP8. In order to validate candidate variants, we performed Sanger sequencing. As the results, we could confirm 5 de novo missense mutations from 5 different genes including; AKNA (AT-hook transcription factor), CELSR3 (cadherin, EGF LAG seven-pass G-type receptor 3), KCTD9 (potassium channel tetramerization domain containing 9), MYH14 (myosin, heavy chain 14, non-muscle) and TCTE1 (t-complex-associated-testis-expressed 1). There was no de novo mutation shared in more than two probands.
Conclusions: In the WES of family samples of ASD, we observed 5 de novo missense mutations from 5 different genes, including AKNA, CELSR3, KCTD9, MYH14 and TCTE1. None of the mutations reported as possible causative genetic mutation in ASD in the previous reports. Of those, CELSR3 is known as playing an important role in planar cell polarity and brain development and maintenance, as well as hippocampal maturation and connectivity (Boutin et al., 2012; Feng et al., 2012). MYH14 has been reported as related with hearing loss, cleft lip and myopathy in human subjects (Choi et al., 2011, Yang et al., 2005).