Fig. 1

CNV analysis strategy. CNV detection in Qataris was assessed at two tiers. First, CNVs were called in 100 individuals using two algorithms each, on two primary input datasets: genotyping array (OMNI2.5 M) and next-generation whole genome sequencing reads (Illumina PE 100 bp, Mean Depth: 37X). A size cut-off of at least 5 consecutive probes for genotyping data and at least 5 consecutive windows for whole genome sequencing data was used to increase specificity (see Methods). Three samples with an unusually high number of CNVs were first removed from the population (see Additional file 1: Figure S1). In the second step, high-quality CNVs from the remaining 97 subjects called by all 4 platforms were distributed into 97 individual files. CNVs were first compared intra-individuals and retained if observed by more than one algorithm. If no overlap was detected within the individual, the CNV was compared inter-individuals to detect a second occurrence in the remaining 97 individuals. CNVs observed only once in the entire sample were discarded. CNVs passing these filters were merged across the population to generate population level CNV regions (CNVRs), which were taken into the detailed analysis steps. *Denotes data was provided as-is from proprietary Illumina Genome Network sequencing pipeline without the ability of the user to alter parameters