Exome Sequencing of Large Multiethnic Population of Moyamoya Patients

We report a significant advance in Moyamoya disease (MMD) research through our unique access to a large MMD patient population combined with high depth exome sequencing and bioinformatics. We confirm a reported RNF213 founder mutation (FM), provide evidence for its ethnic specificity and identify novel variants in other genes that associate with Caucasian and non-RNF213 FM MMD. This work has a broader impact in vascular research by highlighting novel genetic aspects of cerebrovascular disease. Moyamoya Disease (MMD) is a rare cerebrovascular disorder characterized by unilateral or bilateral progressive stenosis or occlusion of the internal carotid artery, with frequent involvement of the anterior cerebral artery and the middle cerebral artery (MCA). Fragile collateral vessels subsequently develop and are particularly prone to hemorrhage. MMD patients are usually diagnosed angiographically following presentation of developmental delays, seizures, migraines, ischemia, and/or hemorrhage. The most common approach to re-establish normal blood flow is bypass surgery. While some progress has been made in identifying disease-associated genes, the underlying disease biology is not well understood. Moyamoya disease (MMD) is a rare cerebrovascular disease. Our unique access to a large MMD patient population combined with high depth exome sequencing and bioinformatics has led to a significant advance in the field. We confirm a reported RNF213 founder mutation (FM), provide evidence for its ethnic specificity and identify novel variants in other genes that associate with Caucasian and non-RNF213 FM MMD. This work has a broader impact in vascular research by highlighting novel genetic aspects of cerebrovascular disease. The 125 ethnically diverse, unrelated MMD patients were matched based on sex and broad ethnic category to 125 controls. Control DNA from the 1000 Genomes project was used to design a control Library by Personalis. 125 controls were selected from this Personalis Control DNA Library for our study. (The 1000 Genomes Project Consortium 2012). Genomic DNA was extracted using the Gentra Puregene kit (Qiagen, Valencia, CA). Libraries were prepared from approximately 3 µg of high quality genomic DNA (50-200 ng/µl) using Illumina TruSeq Genomic DNA High throughput Sample Prep Kits (Illumina, San Diego, CA) and exome enrichment (targeting 62Mb) was accomplished using the TruSeq Exome Target Enrichment kit (Illumina, San Diego, CA), all according to manufacturer's protocols. Target enrichment validation was confirmed by determining the concentration of the library by PicoGreen-based quantitation. Library yields ranged from 100-1000ng of DNA, a portion of which was run on the Bioanalyzer HS DNA chip (Agilent, Santa Clara, CA), with an average size of 300-550nt for DNA fragments. Exome Sequencing. Sequencing was performed using Illumina Hiseq2000 or HiSeq2500 sequencers with single lane, paired-end 2X100bp reads. DNA fragments were generated and amplified using Clonal Single Molecule Array technology (Illumina, San Diego, CA). The sequences were determined using the Clonal Single Molecule Array and Sequencing-by-Synthesis using Illumina's instrumentation and Reversible Terminator Chemistry. Each sequencing lane interrogated the DNA sequences of a pool of 3 individual sample libraries each carrying a unique index. Sequencing reads of at least 2x100bp in length for a total of approximately 8 Gb of sequence data were generated for each sample. ]]> The control and case raw sequencing data were aligned and variants were called simultaneously. Analysis was performed independently in each of the three matched case-control studies to investigate variant and gene associations with MMD. Variants were removed from analysis if the call quality failed our internal quality control (QC) criteria or if more than 30% of the data were missing across samples. Remaining variants were assessed for association with disease status using Fisher's exact test. Effect size was summarized as the OR calculated from the conditional maximum likelihood estimate of a 2x2 contingency table containing alternative and reference allele counts in cases and controls assuming an additive model. Significance testing of the null hypothesis of conditional independence (OR = 1) used a two-tailed test with P values obtained directly from a hypergeometric distribution. In instances where zero cell counts were encountered, either zero or infinite values for the OR were obtained. MMD is a rare, genetically heterogeneous monogenic disease with few disease-causing variants, so we applied additional variant inclusion criteria to reduce false discoveries when reporting MMD-associated variants and genes. Variants were retained if they: (1) were protein-coding SNPs; (2) had a minor allele frequency (MAF) <5% based on ethnic-specific 1000 Genome frequency information; and (3) were enriched in MMD cases. By these criteria, a known MMD allele in the Asian population would be identifiable in previously described studies (RNF213, p.R4810K by the 1000 Genomes has a MAF = 0.001 in all populations and a MAF = 0.002 in Asian population. ]]>