Post-filter CNV counts and relative CNV frequency.

The CNV counts shown here represent the 43,560 candidate CNVs that remain after applying the Post-filter. The first column shows a specific CNV count. The second set of columns show the number of control and case individuals observed at each given CNV count. The center columns show the cumulative count of CNVs, and the last columns show the frequency at which a given number of CNVs are observed in each group of study-participants. A small subset of both case and control samples show highly inflated CNV counts with the highest CNV-counts being 41 and 231 in cases and controls respectively. A review of the individual CNVs in this group revealed that the vast majority of these CNVs are short (less than 20 SNPs), incorrectly called variants of the type CN = 1 and CN = 3. In fact, based on the visual inspection we generally found about 1–3 true CNVs per sample in this group. Using a systematic assessment to identify outliers we classified samples with more than 6 CNVs as outliers prone to increasingly high false-CNV counts.