WHOLE EXOME SEQUENCING OF HEMIPLEGIC MIGRAINE PATIENTS SHOWS AN INCREASED BURDEN OF MISSENSE VARIANTS IN CACNA1H AND CACNA1I GENES 197 9 number of individuals and contains a detailed catalogue of exome-wide genetic variation. Furthermore, gnomAD provided ancestry information. The gnomAD database contains exome variant summary statistics for 56,885 non-Finnish Europeans, with a female-to-male ratio of ~1.27:1, depending on the available genotypes at each specific locus. As HM is a very rare disorder with a prevalence of 0.01%,34 confounding effects due to the presence of HM patients in the control group were deemed to be negligible. Whole exome sequencing and quality control Australian cohort. Following WES, the Ion Torrent Server was used to generate quality metrics, align reads to the Human Genome 19 (Hg19), and the Ion Torrent Variant Caller (TVC) was used to call sequence variants and produce variant calling format (VCF) files. Dutch cohort. Following sequencing, the sequence reads were aligned to the UCSC Genome Browser hg19 reference sequence using the Burrows-Wheeler Alignment tool.35 The generated BAM files were subsequently converted to VCF files using BCFtools.36 Single-variant analysis Prior to performing burden testing all variants were assessed to determine whether there were obvious, high-penetrant disease-causing mutations detected outside of the known HM genes that could cause HM in patients of either cohort. In the absence of such pathogenic mutation, individual missense variants in all the CACNA1x genes were assessed for patients of the Australian cohort. For the Dutch cohort, only those missense variants present in TRAPD-associated CACNA1x genes were investigated. Burden testing Variants pre-processing all cohorts. As VCFs were exported from different platforms, the respective analyses had to be unified. The first step for both cohorts was to normalize VCFs using BCFtools, this ensures that any platform-specific formatting differences are removed, and also expands multiallelic variants.36 VCFs were merged for each cohort using vcftools, and variants with average read depth coverage below 10X were excluded using either BCFtools or the snpEff program.36, 37 For both cohorts, the coding exons of the CACNA1x genes were included with a 5-bp pad on either site of the exon. New VCFs (one merged for each cohort) were annotated with VEP Ensembl.38 For the Dutch cohort as an extra quality control step only those variants with a quality-by-depth (QD) score > 4 were taken forward. Selection of qualifying variants. To determine the number of variants, we selected “qualifying variants” being variants that meet the criteria of inclusion. Only those variants classified (annotated) as missense variants were considered as “qualifying variants”. The number of individuals in the case cohort who carried at least one “qualifying variant” in that gene as well as the total number of
RkJQdWJsaXNoZXIy MTk4NDMw