--outSAMattrRGline is interpreted incorrectly

[byb121]

When using star (version: 2.7.2a) with the command below:

STAR \
--runMode alignReads \
--runThreadN 8 \
--genomeDir star-ref-2.7.2a \
--outSAMattrRGline "ID:2977244" "LB:EM-2_rnas\"|||eq 1\!100\$5739_52200" \
--twopassMode None \
--readFilesIn test_bad_header.1_1.fastq.gz test_bad_header.1_2.fastq.gz \
--outFileNamePrefix /home/ubuntu/tmpStar/ \
--readFilesCommand zcat \
--outSAMtype BAM Unsorted

RG line in the output BAM has header:

$ samtools view -H Aligned.out.bam | grep '@RG'
@RG     ID:2977244      LB:EM-2_rnas    |||eq   1\!100$5739_52200"

but we expect an RG line like this:

@RG     ID:2977244      LB:EM-2_rnas"|||eq 1!100$5739_52200

I believe the "LB" tag string in the command has been escaped correctly:

$ echo "LB:EM-2_rnas\"|||eq 1\!100\$5739_52200"
LB:EM-2_rnas"|||eq 1\!100$5739_52200

I know it's a weird tag value, but every char in the string is acceptable according to SAM/BAM format definitions.

[alexdobin]

Hi Yaobo,

the problem is with having both the " and space in one field, it somehow interferes with the standard c++ text stream input. At the moment, you would need to use samtools reheader to fix the resulting BAM file.

Cheers
Alex