Hi,
I am doing a GWAS analysis of european bison's a dog's SNPs. Right now I am doing multiple data clean up, according to previous studies and what I have learned.
I was thinking, if it is ok to test HWE and discard data that does not fit if I am working with animals, where mating is controlled and inbreeding coefficient is high?
And what about tests for population structure etc.? It is obvious that there is going to be some structure in reintroduced and domestic animals. I don't want to loose any important data and I want to be sure, that I have done everything correctly.
Thank you for your help.
I recommend HWE test to remove the markers that don't fit the equilibrium with threshold p-value of 1e-05. Such threshold is soft enough to leave genotypes that are informative for GWAS but is hard to filter out the markers where the genotypes were determined wrongly. Just in case, what is the size of your population? If it is 20, don't do HWE test. The results will be biased.
Instead of structure analysis I would recommend PCA analysis. It is expected that clouds of case and control groups will be similar on graphics depicting the first three main components pairwise.
Lenka Ungrová If your population is large enough, it is recommended to perform HWE test.
I feel these papers give you some insight into which decision might be right for you:
Article A tutorial on conducting genome-wide association studies: Qu...
Article Departure from Hardy Weinberg Equilibrium and Genotyping Error
Thank you!
Right now I have only small amount of samples, so I won't use HWE. But when the dataset will be larger, I will.
Thank you also for those papers.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics