One of the first procedures that should be implemented in any GWAS QC protocol is checking for potential sample identity problems that typically result from sample handling errors. One of the easiest ways to discover potential sample handling issues that result in mix-ups is by checking the reported sex of each individual against that predicted by the genetic data. The -- check-sex option in PLINK uses X chromosome heterozygosity rates to determine sex empirically, then reports individuals for whom the sex recorded in the pedfile does not match the predicted sex based on genetic data (example output and explanation shown in Table 2). If discrepancies are found (e.g. an individual is recorded being female but appears homozygous for every X chromosome marker), the EMR or any available study questionnaires should be reviewed to make a determination whether there was a sample handling mistake that caused a sample mix-up. Checking X chromosome heterozygosity may also reveal sex chromosome anomalies such as Turner syndrome (females having karyotype XO), Kleinfelter syndrome (males having karyotype XXY), mosaic individuals (XX/XO, XX/XXY), or females with large stretches of loss-of-heterozygosity on the X chromosome who are otherwise phenotypically normal.
