From 636 samples derived from blood or cell line, 622 (97.80 %) passed quality control based on recommended best practices for Axiom arrays by Affymetrix [55]. Several samples were excluded from subsequent analyses; one sample failed to scan, eight had Dish QC < 0.82 due to a sample processing issue, and five had stage one genotype call rates < 97 %. In addition, seven negative controls were processed across the genotyping plates, and all exhibited low separation from the background signal as expected. In passing samples, 582,143 (90.08 %) of the markers on the array passed quality control, using Affymetrix recommended best practices. With default settings of Affymetrix SNPolisher classifications, we removed: 48,083 markers classified as “Other”; 7014 classified as “CallRateBelowThreshold”; 1765 classified as “Off-target variant (OTV)”; 181 classified as “Hemizygous”; and 7061 classified as “MonoHighResolution” that also had previously failed Affymetrix’s internal validation process (see Table 3).Table 3SNP classifications and recommendations using 622 passing samples derived from blood or cell lineSNPolisher ConversionTypePreviously validated markersPreviously failed-validation markersDe novo markersTotal SNPsRecommendation to keep for analysisPolyHighResolution441120 (68.26 %)17595 (2.72 %)7547 (1.17 %)466262
