borgwardtlab / ardiss Goto Github PK

Hi, I've installed the package and have attempted to run the example data but running into an issue:

(ardissvenv) (base) [tutert@rescomp2 demo_data]$ ardiss --typed_zscores insomnia_typed_zscores --reference_genotypes genotypes.npy --markers markers.txt 2021-09-07 09:52:28.432013: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory 2021-09-07 09:52:28.432051: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine. Traceback (most recent call last): File "/gpfs3/well/mcvean/mtutert/sumstat_imputation_paper/ARDISS/ardissvenv/bin/ardiss", line 5, in <module> from ardiss import ardiss_console File "/gpfs3/well/mcvean/mtutert/sumstat_imputation_paper/ARDISS/ardissvenv/lib/python3.8/site-packages/ardiss/__init__.py", line 1, in <module> from .impute_ard import impute_ard File "/gpfs3/well/mcvean/mtutert/sumstat_imputation_paper/ARDISS/ardissvenv/lib/python3.8/site-packages/ardiss/impute_ard.py", line 13, in <module> from .ard_model import GPflowARD, load_ard_weights, scale_with_weights File "/gpfs3/well/mcvean/mtutert/sumstat_imputation_paper/ARDISS/ardissvenv/lib/python3.8/site-packages/ardiss/ard_model.py", line 14, in <module> class GPflowARD(object): File "/gpfs3/well/mcvean/mtutert/sumstat_imputation_paper/ARDISS/ardissvenv/lib/python3.8/site-packages/ardiss/ard_model.py", line 20, in GPflowARD optimizer=gpflow.train.RMSPropOptimizer(0.1, momentum=0.01), AttributeError: module 'gpflow' has no attribute 'train'

Would someone be able to diagnose what is going wrong here?

Hi,
Thank you for making such a wonderful and easy to use software for summary. I have installed and successfully imputed the summary stats using ardiss. I have question about filtering the SNPs. I have summary stats like this:
SNP_id SNP_pos Ref_allele Alt_allele Z-score Var rs61733845 1118275 C T 2.38179318243551 1.0 rs112838421 1118491 A G -0.2858649438410339 0.0 rs113531911 1118560 C T -0.2858649438410339 0.0 rs111290899 1118578 G A 1.5024951548808936 0.0 rs12097131 1118728 C G 0.08270770411550643 0.0

Question : Can I use any column to filter high quality imputed SNPs? Something like real genotyping imputation (INFO > 0.8) or I can use all the SNPs for meta-analysis with other GWAS.
Thanking you advance
Best
Yask

Hi,

When I try to decompress any of the reference files, I get a "gzip: stdin: unexpected end of file" error. I tried downloading the files again and got the same error. I think the files that were uploaded are corrupt, but maybe I am mistaken.

Also the data.tar.gz file with example data I think is in .tar format and is not .gzipped, I had to rename it to remove the .gz extension before I could untar it

bw
Adam

Hi,

Thank you for your tool.
Once the imputation is completed, how can I get the beta and SE of each unmeasured marker?
Thank you

Best,
Audrey

Implement the experiment class rather than relying on impute_ard scripts.

Test gpu flag on the GPU server for the command-line script