Step 1 error about saige HOT 10 CLOSED

Hello Harold,

Could you please upload the original log file? Also can you try LOCO=FALSE to see if the error is because of the LOCO option?

Thank you,
Wei

from saige.

Thanks a lot Wei, here is the original log file for the running.
test_SAIGE_UKBB_step1_v3.log
Let me know if you get it.

I will re-run the analyses by changing the option LOCO=FALSE and also update you again soon.

Thanks a lot!

Harold

from saige.

Hi Harold,

Thank you very much for providing the log file. I'm so sorry I was not able to figure out the possible reason that caused the error at this time. Did you try running the example data to see whether they can be run successfully?

Thanks,
Wei

from saige.

Yes, I did, right after I post my question, I did a test on your example data, it worked, all the results are fine, there is no error from the logs. I have not tried the LOCO=FALSE yet, I will do that soon.

Thanks,

Harold

from saige.

from saige.

Thanks for your suggestions, I will add another test run for PC1-4 as well. I started the one with LOCO=FALSE, it is not done yet.

Harold

from saige.

Hi Wei,

The LOCO=FALSE was done, still encounter the same issue with the computation, here is the log file.
The model with PC1-4 is still runn

test_SAIGE_UKBB_step1_falseLOCO.log

ing.

Harold

from saige.

Hi Wei,

This is quite frustrating that the PC1-4 run is also not successful, it showed similar error message, do you have solution to run chromosome by chromosome instead of all the 22 chromosomes at the same time?

Here is the log file.

Thanks a lot,
test_SAIGE_UKBB_step1_trueLOCO_PC1_PC4.log

Harold

from saige.

Hi Harold,

I'm sorry I couldn't figure out the issue of your jobs. The software has evolved a lot since then. Please reopen the issue if you still encounter this issue with the latest version 0.36.3.1

Thanks!
Wei

from saige.

Hi Wei,
May I ask what "colnames(data.new) is Y minus1" stands for here? minus1 has its own coefficient too.
Thank you
Fedik

from saige.