A total of 636 purified DNA samples (including 27 replicates) derived from blood or cell line were processed on the Smokescreen Genotyping Array at RUCDR Infinite Biologics (Piscataway, NJ) according to manufacturer instructions. Raw data consisted of one CEL file per sample (except for one sample that failed to scan). Following best practices guidelines set forth by the manufacturer [55], the Affymetrix Power Tools (APT) v1.17.0 software was used to process raw data. Dish QC (DQC) values were generated and used to remove samples with DQC < 0.82. As part of stage 1 genotyping, 20,000 probesets previously validated by the manufacturer were used to cluster genotypes and remove samples with stage 1 call rate <97 %. Plate pass rate and average stage 1 call rate per plate were calculated and reviewed to determine if any plates should be excluded from further analysis. Remaining samples were genotyped for all probesets (stage 2 genotyping), using APT. The Affymetrix SNPolisher v1.5.2 software was used to classify probeset quality and determine the best probeset for each marker. Markers whose best probeset classified as “Other”,
