We would like to compute variance components with multiple covariance matrices, using
./gemma-0.98.1-linux-static -p y.pheno.txt -n 1 -mk matrix.txt -vc 1 -o out.txt
For creating matrix.txt we used:
printf './Kin.cXX.txt\n./H.cXX.txt\n' > matrix.txt
However error appears as '**** FAILED: number of rows in the kinship file does not match the number of individuals.'.
Separately using Kin.cXX.txt and H.cXX.txt will work, but multiple variance components estimation is desired.
In addition, there seems to be no option of including covariates into the command for variance component estimation; by passing only the regressed residuals will produce wrong results due to the wrong covariance matrix of error term assumed (should be I-W(W^TW)^{-1}W^T, rather than I-11^T/n as implemented).
We would really appreciate it if these issues could be solved.
We would like to compute variance components with multiple covariance matrices, using ./gemma-0.98.1-linux-static -p y.pheno.txt -n 1 -mk matrix.txt -vc 1 -o out.txt For creating matrix.txt we used: printf './Kin.cXX.txt\n./H.cXX.txt\n' > matrix.txt However error appears as '**** FAILED: number of rows in the kinship file does not match the number of individuals.'. Separately using Kin.cXX.txt and H.cXX.txt will work, but multiple variance components estimation is desired. In addition, there seems to be no option of including covariates into the command for variance component estimation; by passing only the regressed residuals will produce wrong results due to the wrong covariance matrix of error term assumed (should be I-W(W^TW)^{-1}W^T, rather than I-11^T/n as implemented).
We would really appreciate it if these issues could be solved.