GATK gCNV doesn't recognize BAM index input which isn't located in the same origin folder as BAM file #7487
Comments
|
@mwalker174 ^^ |
|
@droazen isn't this a simple case of using another command line argument to specify the index path? And if that's necessary for DRS files, won't it be the same for all tools? |
|
@ldgauthier It's true that this issue will affect other tools/pipelines, but to address this specific issue the gCNV WDL will need to be patched to either copy the |
|
Any reason we wouldn't want to always use --read-index in WDLs, even if it's named as expected? |
|
@seunghun23 I just merged a fix into the master. Closed by #7518. |
|
Hi @asmirnov239 is it also possible to add this change to gatk/mutect2-gatk4? I am having the same problem there. Thanks! |
|
@s-hoyt I am looking at the mutect2 WDL (https://github.com/broadinstitute/gatk/blob/master/scripts/mutect2_wdl/mutect2.wdl) and it seems like there is an option to specify BAM indices as a workflow level input. Are there any specific tasks that you had trouble with? |
|
@asmirnov239 I'm getting the same error on Terra when bam files and their indices are not in the same location. I'm running https://github.com/broadinstitute/gatk/blob/4.1.7.0/scripts/mutect2_wdl/mutect2_pon.wdl. The workflow input for bam index seems to be just a placeholder, if you follow the variables along the wdl it's never actually used. |
|
Seems to be an easy fix similar to #7518, adding |
Hi,
I'm trying to run "cnv_germline_cohort_workflow" from this workspace (https://app.terra.bio/#workspaces/help-gatk/Germline-CNVs-GATK4), and the workflow is keep failing at the "CollectCounts" step with the following error
in multiple shards
A USER ERROR has occurred: Traversal by intervals was requested but some input files are not indexed.
Please index all input files:
samtools index /cromwell_root/dg.4DFC_3615e55e-6aa3-43e7-8d7b-6f2824071971/C345.TCGA-A3-3373-11A-01D-1421-08.5.bam
This was weird since I set a correct index file as an input, but after some investigation, I realized that
the error seemed to be occurring when the bucket path where BAM index was located was different from that of BAM file
For example, if you look at this log below,
Attempting to download gs://gdc-tcga-phs000178-controlled/KIRC/DNA/WXS/BI/ILLUMINA/C345.TCGA-A3-3373-11A-01D-1421-08.5.bam to /cromwell_root/dg.4DFC_3615e55e-6aa3-43e7-8d7b-6f2824071971/C345.TCGA-A3-3373-11A-01D-1421-08.5.bam Successfully activated service account; Will continue with download. Activated service account credentials for: [kd5mqbpsed8lzz0kyz9tvkht-3274@dcf-prod.iam.gserviceaccount.com] Download complete! 2021/09/29 15:46:14 Localizing input drs://dg.4DFC:ab4d57fa-bfff-4a48-bd96-f2866ecfe0e1 -> /cromwell_root/dg.4DFC_ab4d57fa-bfff-4a48-bd96-f2866ecfe0e1/C345.TCGA-A3-3373-11A-01D-1421-08.5.bam.bai Requester Pays project ID is Some(vanallen-firecloud-nih) Attempting to download gs://gdc-tcga-phs000178-controlled/KIRC/DNA/WXS/BI/ILLUMINA/C345.TCGA-A3-3373-11A-01D-1421-08.5.bam.bai to /cromwell_root/dg.4DFC_ab4d57fa-bfff-4a48-bd96-f2866ecfe0e1/C345.TCGA-A3-3373-11A-01D-1421-08.5.bam.bai Successfully activated service account; Will continue with download. Activated service account credentials for: [kd5mqbpsed8lzz0kyz9tvkht-3274@dcf-prod.iam.gserviceaccount.com] Download complete!C345.TCGA-A3-3373-11A-01D-1421-08.5.bam and C345.TCGA-A3-3373-11A-01D-1421-08.5.bam.bai
were successfully downloaded, but since these TCGA files use DRS URI, they were copied to two separate cromwell folders
/cromwell_root/dg.4DFC_3615e55e-6aa3-43e7-8d7b-6f2824071971/C345.TCGA-A3-3373-11A-01D-1421-08.5.bam
/cromwell_root/dg.4DFC_ab4d57fa-bfff-4a48-bd96-f2866ecfe0e1/C345.TCGA-A3-3373-11A-01D-1421-08.5.bam.bai
And GATK doesn't seem to recognize BAM index when it is not inside a same folder.
Could you maybe add symlink for the BAM and BAI files in the WDL script?
Thanks,
Seunghun
The text was updated successfully, but these errors were encountered: