ENCODE-DCC / atac-seq-pipeline

ENCODE ATAC-seq pipeline
MIT License
385 stars 172 forks source link

build custome genome index using script "build_genome_data.sh" #273

Open yysheep2020 opened 4 years ago

yysheep2020 commented 4 years ago

Hi dear Jin,

when I run the script "build_genome_data.sh", it showed the error:

**_Total time for backward call to driver() for mirror index: 00:34:14

tar: unrecognized option '--sort=name' Try tar --help' ortar --usage' for more information._**

SO I think this error is rising because the command in script: tar cvf ${REF_FA_PREFIX}.tar ${REF_FA_PREFIX}.*.bt2 --sort=name --owner=root:0 --group=root:0 --mtime="UTC 2019-01-01"

Could you help me to check this?

yysheep2020 commented 4 years ago

Hi Jin,

An updated The stanford scg cluster has tar(GNU tar) version 1.26. But command "--sort=name" introduced to 1.28 version. could you help me to modify your code?

leepc12 commented 4 years ago
$ source activate encode-atac-seq-pipeline
$ conda install tar==1.29 -c conda-forge
leepc12 commented 4 years ago

Closing this. Activating pipeline's Conda env should work since tar is included in the requirements so latest GUN tar should be installed inside the env.

leepc12 commented 4 years ago

Reopened. Please let me know if the Conda solution works.

batzza commented 1 year ago

Not the orignal poster but I faced the same issue/error message. I used the conda tar update solution proposed above and it fixed the problem.