Post by rr on Dec 19, 2016 0:12:35 GMT
Hi!
I have the following kind of data:
2434 samples: 1231 cases, 1203 controls
around 555K SNPs
After performing the QC approach similar to Lee et al (2011), the final dataset had almost 499K SNPs and 1947 samples (990 cases and 957 controls).
I performed REML estimation of the variance explained by the SNPs and the result was V(G)/Vp = 0.999998 with 0.000002 SE.
Is it possible to have such very high estimate (and with very little SE) or is there some kind of overfitting?
I have the following kind of data:
2434 samples: 1231 cases, 1203 controls
around 555K SNPs
After performing the QC approach similar to Lee et al (2011), the final dataset had almost 499K SNPs and 1947 samples (990 cases and 957 controls).
I performed REML estimation of the variance explained by the SNPs and the result was V(G)/Vp = 0.999998 with 0.000002 SE.
Is it possible to have such very high estimate (and with very little SE) or is there some kind of overfitting?