After refining the stringency for armadillo repeat enrichment in the discovery data set, it is clear that armadillo repeats are also recurrently enriched in our data. As described earlier, we limited the discovery analysis to only the validated probands/samples, but refined the enrichment scoring so that only regions within armadillo repeats exhibiting greater than 50% sequence homology (including non-coding regions) were considered as enrichment signals. We filtered on our validated probands and did not restrict the analysis to filtered genomes. Enriched regions were characterized as having a false discovery rate (FDR) < 5%, and p-value < 5 x 10−6 (Fig. 7a).
We conclude that the armadillo gene cluster generally demonstrates significant enrichment of DNA sequence in our data validating previously described armadillo transcript variations in ARVC. This can be further supported by the recent observation of a novel armadillo repeat expansion in an ARVC family pedigree. The armadillo gene cluster is a very promising candidate for further investigation into the genetic etiology of ARVC and related cardiac disorders.
We next performed the same regional analysis of enrichment in the validation data set, using the same window size and MAF restrictions (Fig. 3b). While there was no clear enrichment in the discovery data set, there were two regions in particular that were statistically enriched in this validation data set (FDR 0.01). The odds of DNA sequence matching to these two regions of the armadillo gene cluster, is 1.08 x 10−3 and 1.19 x 10−3, respectively. [See Supplemental Table 1 for the detailed definition of enriched armadillo regulatory regions defined by regr pipeline.]
We then focused on the limited genomic regions of armadillo repeats above the observed Q5 threshold of regions with more than 50% sequence similarity (see Figure 3a). We next performed a regional analysis using the same window size and MAF restrictions in the validation data set, and filtered on our validated probands/samples (FDR 0. 7211a4ac4a