|
Post by weihuameng on Nov 25, 2014 15:52:56 GMT
I am calculating narrow sense heritability for a cases-control GWAS study using genotyped SNPs. qualitative trait.
My question is why the heritability of the overall dataset (0.117) is lower than female dataset (0.167) and male dataset (0.28)? I guess it has something to do with the overalll sample size is bigger ? But can anyone provide more scientific explanations? Thank you a lot.
For the over all dataset: h2= 0.117 Source Variance SE V(G) 0.019111 0.009623 V(e) 0.144592 0.009931 Vp 0.163704 0.003420 V(G)/Vp 0.116743 0.058626
For female dataset:h2=0.167 Source Variance SE V(G) 0.032089 0.026904 V(e) 0.160390 0.027125 Vp 0.192480 0.006345 V(G)/Vp 0.166716 0.139428
For male only dataset:h2 = 0.285 Source Variance SE V(G) 0.040967 0.014298 V(e) 0.102580 0.014207 Vp 0.143547 0.003887 V(G)/Vp 0.285393 0.098667
|
|
|
Post by Jian Yang on Nov 26, 2014 0:28:01 GMT
I think it might be just due to sampling. The SEs are large. The difference in estimate between males and females is not significant. diff = 0.285 - 0.167 = 0.12 and SE(diff) = sqrt(0.139^2 + 0.099^2) = 0.17.
|
|
|
Post by weihua meng on Nov 26, 2014 10:22:19 GMT
So it is not my calculation mistake. It is just caused by the sample difference males and females? People (my boss) might wonder why the overall heritability is not a value between 0.167-0.285. but outside as 0.117
|
|
|
Post by Jian Yang on Dec 2, 2014 22:29:00 GMT
re 1) There could also be a real difference between sexes but the sample size is not large enough to detect that.
re 2) The overall estimate takes into account the comparisons between sexes.
|
|