paperKB
coga / coga-kb
Processing
Help
Sign in

Chunk #2 — DETERMINING PARENTAL ORIGIN

Source
Parental origin of sequence variants associated with complex diseases.
Embedded
yes

Text

After filtering based on various quality and yield criteria, 289,658 autosomal markers and 8,411 markers on chromosome X were used. Excluding those with no parent listed in the genealogy database or with less than 98% genotyping yield, 38,167 individuals, a majority typed with Illumia HumanHap300/CNV370 bead chips (Supplementary Information), were processed. For these individuals, 97.8% of the heterozygous genotypes were long range phased, 99.8% of these had parental origin determined. Overall, 3,841,331,873 heterozygous genotypes, 97.7% of those called, had parental origin assigned. The data includes 2,879 typed trios. To empirically evaluate the accuracy of our method, a run was performed with the data of parents in these trios removed when determining parental origin. For 231,585,437 heterozygous genotypes in the probands/offspring, parental origin was determined both by our method and by the trio data directly, with 500,330 discrepancies, an error rate of 500,300/231,585,437 = 0.22%. Since the trios tested have passed heritability checks in pre-processing, the error rate for individuals with fewer than two parents genotyped is probably higher. Still, the overall error rate is likely below 0.4% (Supplementary Information).