Dear Christopher,
I am currently am preparing to run a Transmission Disequilibrium Test (TDT).
My dataset has a specific structure: it consists entirely of mother-child duos (father information is completely missing).
While trying to understand how PLINK handles missing parents during these tests, I reviewed the PLINK 1.9 source code (specifically family.c). I noticed that the get_trios_and_families function accepts an include_duos parameter. However, when this function is called within the tdt(), dfam(), and qfam() functions, the parameter is hardcoded to 0 (e.g., retval = get_trios_and_families(..., 0, multigen);). As a result, my duo families are entirely skipped during the analysis.
I have two questions regarding this:
Could you kindly explain the statistical or algorithmic reasoning behind forcing the exclusion of duos in these family-based association tests? I suspect it might be related to the difficulty of properly testing the null hypothesis of a 50:50 transmission rate without both parents, but I would greatly appreciate your expert insight.
Given my data structure, is writing a custom script (e.g., in R or Python) to manually count the transmitted/untransmitted alleles from heterozygous mothers and performing a binomial test the only statistically sound workaround? Or is there a specific method or flag in PLINK 1.9 (or PLINK 2.0) that I might have missed that can handle this duo-only scenario?
Thank you very much for your time and for maintaining such an invaluable tool.
Sincerely,
Yujin
--
You received this message because you are subscribed to the Google Groups "plink2-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plink2-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/plink2-users/85c2d51d-d8ec-4f32-81f0-cd9cb46619d4n%40googlegroups.com.