Population Genotyping

During year 2, Leducq investigators deployed a custom designed Illumina "Golden Gate" system to interrogate 1,536 SNP sites. The goal of this experiment is to identify common variants in very high likelihood candidate genes modulating SCD risk. Thus, the vast majority of SNPs selected for assay are known to have minor allele frequencies >5%, are in non-coding regions, tag imputed haplotype blocks, and are presumed to represent variations in transcriptional regulation of the genes. SNPs were then examined in the Illumina informatics pipeline, and those with predicted high design scores included; those that did not pass the design stage were replaced were possible by other SNPs tagging the same haplotype blocks. The platform also included ~20 non-synonymous SNPs (i.e. those that change the encoded amino acid) with lower minor allele frequencies (1-5%) but known to alter ion channel as well as 103 SNPs assayed to ensure that the populations interrogated were drawn from a similar background ("genomic controls"). The final list is shown in table 1.

The genotyping was performed over the summer of 2007 on an Illumina Beadstation 500G by staff directed by the Munich site of this Network. The experiment analyzed 4600 samples, with an average-call rate per sample of 97.1%. Over 7 million genotypes were generated in the populations listed in table 2. Analysis is underway.

Drug-induced long QT: The drug-induced long QT syndrome set - generated by 6 Network centers - is the largest such collection in the world, and thus represents a unique opportunity to explore genomic determinants of this serious adverse drug reaction. We define this syndrome as the development of wither a QT interval >600 msec and/or torsades de pointes on exposure to a culprit drug, with a QT of <500 msec on drug withdrawal or at baseline; 275/329 cases had torsades de pointes. For this experiment, three control groups will be used. The first is patients exposed to QT-prolonging antiarrhythmics without developing QT interval prolongation (<500 msec). The second is patients with KCNH2/HERG mutations. Since the vast majority of drugs causing torsades are HERG/IKr blockers, we hypothesize that the polymorphisms determining variability in the extent of QT prolongation by drugs will be the same as those determining variability in the QT interval when IKr is rescued by a genetic mutation. The third control group is a highly-phenotyped set of Caucasian population controls ascertained in southern Germany, from the KORA collection.