statgen
Emmax test
Hello!
I am running emmax gene-based CMC test on the sample 47,000 participants. The input vcf file includes 1824 variants. The number of individuals is the same kinship, vcf and ped file. However, the analysis is taking more than 2 days. So far no file with association results has been generated, only Makefile, .phe, .ind, .grp , .cov and .reml files and a large .eigr.R file. I am using epacts version UKBB.chr1.22.emmaxCMC.AC.eigR Is there any reason why the analysis is taking so long? Thank you for your help!
Is the timestamp on the eigr file recent? Can you check that you are not low on RAM and using swap?
the eigr was created on 07-04 18:19. RAM: MemTotal: 196533944 kB MemFree: 36321224 kB MemAvailable: 159619784 kB
Sorry, I was actually looking for the modification timestamp. I'm trying assess whether the reml step is still running and making progress. Is there anything informative in the stderr output?
I have actually cancelled the jog for that file, because it was taking too long. I ran the analysis on chromosome 17, and got and error: NOTICE - Reading eigenvectors NOTICE - Allocating a size 18446744071655423224 bytes
terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc make: *** [chr17.emmaxCMC.AC.no.cov.0.epacts] Aborted
Hmm... that is way too much memory for 47,000 individuals. Seems like a bug.
Can you try the latest pre-release https://github.com/statgen/EPACTS/releases or the latest from the develop branch?
I don't think EMMAX works for >20K samples..
Hyun.
Hyun Min Kang, Ph.D. Associate Professor of Biostatistics University of Michigan, Ann Arbor Email : @.***
On Thu, Apr 8, 2021 at 3:41 PM Jonathon LeFaive @.***> wrote:
Hmm... that is way too much memory for 47,000 individuals. Seems like a bug.
Can you try the latest pre-release https://github.com/statgen/EPACTS/releases or the latest from the develop branch?
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/statgen/EPACTS/issues/30#issuecomment-816109193, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABPY5OLRHKGACZNLR5EMSNDTHYBGPANCNFSM42RO5TZQ .
@LivUllmann just curious if 3.4.2 release was able to solve your problem? I am working with a vcf file with 61000 samples and 41000 variants. its been 22 days emmaxCMC is running but not output yet. it is consuming only 1 cpu to 100% and 170 gb memory.
I don't think EMMAX works for >20K samples.. Hyun. ----------------------------------------------------- Hyun Min Kang, Ph.D. Associate Professor of Biostatistics University of Michigan, Ann Arbor Email : @.*** … On Thu, Apr 8, 2021 at 3:41 PM Jonathon LeFaive @.***> wrote: Hmm... that is way too much memory for 47,000 individuals. Seems like a bug. Can you try the latest pre-release https://github.com/statgen/EPACTS/releases or the latest from the develop branch? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#30 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABPY5OLRHKGACZNLR5EMSNDTHYBGPANCNFSM42RO5TZQ .
Dear Prof Dr. Kang, just wondering if there is a way to handle 61000 samples in EPACTS?