Post by justgettinstarted on Sept 30, 2015 20:34:20 GMT
I'm the thinking about the possibility of using the grm to determine cryptic relatedness in GWAS sample. There are several reasons (not the least of which is processing time) why i dont like the pairwise comparison produced by the PLINK --genome command. A colleague mentioned that he used GCTA --grm-cutoff 0.025 command to extract unrelated individuals. For my purposes i don't need to remove related individuals but would like to assign them to family groups (FIDs) and then incorporate them as a random effects variable in mixed model regressions. My questions are as follows.
Is this a terrible idea?
Can the grm matrix produced using GCTA even be used to assess cryptic relatedness?
What are the expected Ajk values for 1) duplicate samples/monozygotic twins, 2) 1st degree relatives (parents/offspring, siblings), 3) 2nd degree relatives (grandparents, aunts/uncles, 1/2 siblings), 4) 3rd degree relatives (eg cousins) etc...?
Appreciate your thoughts on this
Is this a terrible idea?
Can the grm matrix produced using GCTA even be used to assess cryptic relatedness?
What are the expected Ajk values for 1) duplicate samples/monozygotic twins, 2) 1st degree relatives (parents/offspring, siblings), 3) 2nd degree relatives (grandparents, aunts/uncles, 1/2 siblings), 4) 3rd degree relatives (eg cousins) etc...?
Appreciate your thoughts on this
