Rare variants in ischemic stroke: an exome pilot study

Abstract

The genetic architecture of ischemic stroke is complex and is likely to include rare or low frequency variants with high penetrance and large effect sizes. Such variants are likely to provide important insights into disease pathogenesis compared to common variants with small effect sizes. Because a significant portion of human functional variation may derive from the protein-coding portion of genes we undertook a pilot study to identify variation across the human exome (i.e., the coding exons across the entire human genome) in 10 ischemic stroke cases. Our efforts focused on evaluating the feasibility and identifying the difficulties in this type of research as it applies to ischemic stroke. The cases included 8 African-Americans and 2 Caucasians selected on the basis of similar stroke subtypes and by implementing a case selection algorithm that emphasized the genetic contribution of stroke risk. Following construction of paired-end sequencing libraries, all predicted human exons in each sample were captured and sequenced. Sequencing generated an average of 25.5 million read pairs (75 bp×2) and 3.8 Gbp per sample. After passing quality filters, screening the exomes against dbSNP demonstrated an average of 2839 novel SNPs among African-Americans and 1105 among Caucasians. In an aggregate analysis, 48 genes were identified to have at least one rare variant across all stroke cases. One gene, CSN3, identified by screening our prior GWAS results in conjunction with our exome results, was found to contain an interesting coding polymorphism as well as containing excess rare variation as compared with the other genes evaluated. In conclusion, while rare coding variants may predispose to the risk of ischemic stroke, this fact has yet to be definitively proven. Our study demonstrates the complexities of such research and highlights that while exome data can be obtained, the optimal analytical methods have yet to be determined.

Figure 1. Summary of the pipeline steps.

Figure 2. SNPs in the 10 cases exomes vs. 3 control populations.

Figure 3. Compound heterozygotes in exome data for genes in which every case had a least two novel variants in the same gene isoform, only 6 genes (9 total isoforms) satisfied this criterion.

The figure illustrates the 9 isoforms (left y-axis) per sample (x-axis) by variant type (left y-axis). Notably, two of the CTPB2 isoforms (NM_001329 and NM_ 001083914) were seen to have a non-sense codon occurring in all 10 samples.

Figure 4. This image shows the start region for CTPB2 gene (chromosome 10) in the Integrated Genomics Viewer.

The top three tracks show aligned sequence coverage for three of the pilot exomes and demonstrate the T to A substitution resulting in the nonsense codon as compared to the two bottom tracks showing the annotated RefSeq and codon positions.