Pe100 sequencing

Sequence data were generated by the Beijing Genomics Institute (BGI), Hong Kong, China. For library production the purified DNAs were digested with ApeKI. The library was subjected to Illumina PE100 sequencing. The raw sequence reads were submitted to the Sequence Read Archive (SRA, NCBI) under accession number PRJNA655397.

In order to determine which SNPs on the OMNI 2.5 M array adequately tag CNVRs observed in the Qatari population, we selected CNV deletions that were observed at least 4 times in the population of 97 (MAF ~2 %). The CNV genotypes were converted to binary format along with all genotypes for use by PLINK [60 (link)]. The PLINK-pairwise-ld command was used to calculate Pearson’s correlation between the genotype of each CNV and all neighboring SNPs ± 500 kb of the CNV’s breakpoints across 97 individuals. For each CNV, the closest SNP with the highest r² value within the 500 kb window was identified.
Similarly, tagging SNPs from the whole genome sequencing data were identified for these same deletions. Whole genome sequencing data from all 97 individuals were obtained from Illumina PE 100 Sequencing (as described above). These were pruned for variants within 500 kb upstream and downstream of each CNV, and then plink was used as before to determine SNPs with the highest Pearson correlation (maximum r² value) within these windows.