MergeBamAlignment conflicting headers between mapped and unmapped

[GATKSupportTeam]

MergeBamAlignment should throw an error message when the BAM files have conflicting @rg tags in the headers.

This request was created from a contribution made by TA on March 04, 2022 15:13 UTC.

Link: https://gatk.broadinstitute.org/hc/en-us/community/posts/4532746666523-MergeBamAlignment-conflicting-headers-between-mapped-and-unmapped

--

Hi,

I wanted to ask how MergeBamAlignment deals with conflicting header information. Specifically if the unmapped BAM and mapped BAM have different @rg tags in the header then how does it resolve? I do not see this in documentation.

My issue is that I add @rg tag information during alignment with BWA-MEM. But when I use mergebamalignment while the reads retain the @rg ID the header changes the @rgid to the unmapped BAM. Therefore the read group IDs of the actual reads is different than the header. I have three screen shots below of this. 

I don't see any documentation of how to deal with conflicting headers - I know a workaround is to change the readgroups with AddOrReplace read groups after MergeBamAlignmnet but I would rather do this at the level of MergeBamAlignment

This question is similar but doesn't say how it resolves this conflict: https://gatk.broadinstitute.org/hc/en-us/community/posts/4406765311515-Read-group-info-in-mapped-and-unmapped-bam-files?input\_string=MergeBamAlignment%20conflicting%20headers%20between%20mapped%20and%20unmapped

Thank you!

After merge bam alignment RG

BWA-MEM mapped RG

Unmapped read group

My version of GATK is 4.2.1.0

My MergeBamAlignment command was

gatk MergeBamAlignment Aligned=sorted_mapped_group"$read_batch_num".bam UNMAPPED=sorted_unmapped_group"$read_batch_num".bam o=bam_aligned_"$read_batch_num".bam R=~/evronylab/reference-files/hg38-gatk/Homo_sapiens_assembly38.fasta

(created from Zendesk ticket #275269)
gz#275269