Post by gracekatherine on Nov 6, 2015 15:34:30 GMT
Hello,
I am currently working on a project with data from two different consent groups. There are no overlapping individuals between the two groups, and I would like to run a REML analysis on a single combined GRM rather than the two cohort GRMs separately in order to maximize the sample size.
Is it possible to combine these cohorts to create a single GRM file? I tried using --mgrm and encountered the following error:
Analysis started: Fri Nov 6 10:19:20 2015
Options:
--mgrm multi_test.txt
--maf 0.01
--make-grm
--out multi_test
Note: This is a multi-thread program. You could specify the number of threads by the --thread-num option to speed up the computation if there are multiple processors in your machine.
There are 2 GRM file names specified in [multi_test.txt].
Error: no individual is in common in the GRM files.
The command I submitted reads as follows:
gcta64 --mgrm multi_test.txt --maf 0.01 --make-grm --out multi_test
Where multi_test.txt is a text file that includes the names of the two separate GRMs from each cohort.
I also tried running the REML analysis on the two separate GRMs without combining using --mgrm but ran into problems there as well:
Analysis started: Fri Nov 6 10:16:42 2015
Options:
--mgrm multi_test.txt
--pheno BMI.phen
--reml
--out test_BMI
Note: This is a multi-thread program. You could specify the number of threads by the --thread-num option to speed up the computation if there are multiple processors in your machine.
Reading phenotypes from [BMI.phen].
Non-missing phenotypes of 1152 individuals are included from [BMI.phen].
There are 2 GRM file names specified in the file [multi_test.txt].
Reading the GRM from the 1th file ...
Reading IDs of the GRM from [multi_chr.grm.id].
1153 IDs read from [multi_chr.grm.id].
Reading the GRM from [multi_chr.grm.bin].
Pairwise genetic relationships between 1153 individuals are included from [multi_chr.grm.bin].
Reading the GRM from the 2th file ...
Reading IDs of the GRM from [REDACTED PATH].
3727 IDs read from [REDACTED PATH].
Reading the GRM from [REDACTED PATH].
Pairwise genetic relationships between 3727 individuals are included from [REDACTED PATH].
1152 individuals are in common in these files.
Performing REML analysis ... (Note: may take hours depending on sample size).
1152 observations, 1 fixed effect(s), and 3 variance component(s)(including residual variance).
Calculating prior values of variance components by EM-REML ...
Error: the variance-covaraince matrix V is not positive definite.
The command I submitted reads as follows:
gcta64 --mgrm multi_test.txt --pheno BMI.phen --reml --out test_BMI
Thank you!
I am currently working on a project with data from two different consent groups. There are no overlapping individuals between the two groups, and I would like to run a REML analysis on a single combined GRM rather than the two cohort GRMs separately in order to maximize the sample size.
Is it possible to combine these cohorts to create a single GRM file? I tried using --mgrm and encountered the following error:
Analysis started: Fri Nov 6 10:19:20 2015
Options:
--mgrm multi_test.txt
--maf 0.01
--make-grm
--out multi_test
Note: This is a multi-thread program. You could specify the number of threads by the --thread-num option to speed up the computation if there are multiple processors in your machine.
There are 2 GRM file names specified in [multi_test.txt].
Error: no individual is in common in the GRM files.
The command I submitted reads as follows:
gcta64 --mgrm multi_test.txt --maf 0.01 --make-grm --out multi_test
Where multi_test.txt is a text file that includes the names of the two separate GRMs from each cohort.
I also tried running the REML analysis on the two separate GRMs without combining using --mgrm but ran into problems there as well:
Analysis started: Fri Nov 6 10:16:42 2015
Options:
--mgrm multi_test.txt
--pheno BMI.phen
--reml
--out test_BMI
Note: This is a multi-thread program. You could specify the number of threads by the --thread-num option to speed up the computation if there are multiple processors in your machine.
Reading phenotypes from [BMI.phen].
Non-missing phenotypes of 1152 individuals are included from [BMI.phen].
There are 2 GRM file names specified in the file [multi_test.txt].
Reading the GRM from the 1th file ...
Reading IDs of the GRM from [multi_chr.grm.id].
1153 IDs read from [multi_chr.grm.id].
Reading the GRM from [multi_chr.grm.bin].
Pairwise genetic relationships between 1153 individuals are included from [multi_chr.grm.bin].
Reading the GRM from the 2th file ...
Reading IDs of the GRM from [REDACTED PATH].
3727 IDs read from [REDACTED PATH].
Reading the GRM from [REDACTED PATH].
Pairwise genetic relationships between 3727 individuals are included from [REDACTED PATH].
1152 individuals are in common in these files.
Performing REML analysis ... (Note: may take hours depending on sample size).
1152 observations, 1 fixed effect(s), and 3 variance component(s)(including residual variance).
Calculating prior values of variance components by EM-REML ...
Error: the variance-covaraince matrix V is not positive definite.
The command I submitted reads as follows:
gcta64 --mgrm multi_test.txt --pheno BMI.phen --reml --out test_BMI
Thank you!