RG required in BAM in latest branch release_1.0.2?

[wshands]

We are seeing failures like this

2022-04-14 23:15:26 : (BAM_data : 1 of 1) Tabix Bam To Temporary File, 10.091 mins elapsed.

************************************************************
2022-04-14 23:16:02 : ERROR Found in .bamToTmp for (BAM_data : 1 of 1) 
LogFile = ArchRLogs/ArchR-createArrows-15edf0b5b-Date-2022-04-14_Time-23-05-21.log

<simpleError in eval(quote(list(...)), env): object 'RG' not found>

************************************************************

I believe when using a bam file with no Read Group in the BAM header as input to CreateArrowFiles in branch release_1.0.2.

Maybe in commit fe48101 the assumption that a RG is in the header was introduced here (I'm not an experienced R programmer so I may be way off)

At least we don't see this problem installing with an earlier commit, i.e.
devtools::install_github("GreenleafLab/ArchR@63e7883586a013b9a9c264ddb19dea18d9287087", ref="release_1.0.2", repos = BiocManager::repositories())
Would you let us know if this is correct? Thanks!

[rcorces]

BAM file input requires specification of where the cellular barcode is stored. This is done using the bcTag parameter. Whatever is stored in that BAM field is what gets stored as RG. RG is not a required field / there is no assumption that RG is in the header. The RG header is created from bcTag

ArchR/R/CreateArrow.R

Line 1585 in c61dfca

RG = scanChunk$tag[[bcTag]]