Post by jonic on Jun 7, 2018 8:37:59 GMT
Hi,
I'm running GSMR using GCTA 1.91.4 beta3. I'm running several different analyses, and it runs on almost all of them, but for one analysis I get this output:
I know non-inversions are usually due to co-linearities or similar preventing the model from fitting, but I have successfully run this exposure with other outcomes (and with an outcome that is a subset of the failing outcome).
What could be causing this invertibility, and is there a way I can solve it?
Full output is attached in case it is useful!
Thanks,
Joni
I'm running GSMR using GCTA 1.91.4 beta3. I'm running several different analyses, and it runs on almost all of them, but for one analysis I get this output:
Reverse GSMR analysis for exposure #5 and outcome #1 ...
57 index SNPs are obtained from the clumping analysis.
LD clumping with a FDR threshold of 0.00 and a LD r2 threshold of 0.05 ...
After LD clumping, there are 57 SNPs, 0 SNPs removed.
55 index SNPs are retained after the HEIDI-outlier analysis.
Error: The variance-covariance matrix of bxy is not invertible.
An error occurs, please check the options or data
I know non-inversions are usually due to co-linearities or similar preventing the model from fitting, but I have successfully run this exposure with other outcomes (and with an outcome that is a subset of the failing outcome).
What could be causing this invertibility, and is there a way I can solve it?
Full output is attached in case it is useful!
Thanks,
Joni