|
Post by yw on Mar 25, 2014 13:27:06 GMT
Hello, I am looking to calculate the Genetic Correlation between two traits. I have a sample of 1000 individuals, and 900K SNPs. It is said (in gcta.freeforums.net/thread/4/sample-size) that SE is ~300/N with N being the sample size so it appears that I have too few samples. (This is said for heritability, but is this also true for genetic correlation?) Is there any solution to this problem? Would it help If I had more SNPs?
|
|
|
Post by Jian Yang on Mar 26, 2014 11:20:17 GMT
The SE of the estimate of genetic correlation (rG) is function of sample size, h2G of both traits, and rG itself. It won't help if you have more SNPs. Given a sample size of 1000, the SE of rG will be too large to make any meaningful conclusion.
|
|