Error parsing SAM header. @RG line missing SM tag

Question

[kwangchol.mun]

I have a bam file called test.bam with malformed @RG line. @RG line has only one parameter as following:

@RG ID:foo

I want to add the parameters like SM, LB, PL to @RG line, but I have difficulties to do it.

I already tried with samtools as following:

samtools view -H test.bam | sed 's,^@PG.*,@PG\tID:None\tSM:None\tLB:None\tPL:Illumina,g' |  samtools reheader - test.bam > test.rg.bam

However, It gives an error like this:

[E::sam_hdr_error] Missing tab at line 198: "@PGtID:NonetSM:NonetLB:NonetPL:Illumina"

I also tried picard with AddOrReplaceReadGroups function, but it also replied with errors!

java -jar picard.jar AddOrReplaceReadGroups I=test.bam O=test.out.bam RGID=4 RGLB=lib1 RGPL=illumina RGPU=unit1 RGSM=20

Exception in thread "main" htsjdk.samtools.SAMFormatException: Error parsing SAM header. @RG line missing SM tag. Line: @RG ID:foo

Please help!

Thank you in advance

[kwangchol.mun]

Thank you! samtools command does not work but the VALIDATION_STRINGENCY=LENIENT option works! Cheers

Answer 1

[Pierre Lindenbaum]

in picard/htsjdk RG must have another tag than ID, the separator is the tabulation.

try to fix this with:

samtools -h view in.bam | sed 's/^\(@RG\tID:.*\)/\1\tSM:foo/' | samtools view -b -o out.bam -

[Pierre Lindenbaum]

another option: use AddOrReplaceReadGroups with the option VALIDATION_STRINGENCY=LENIENT

[GenoMax]

Sorry about the off-topic comment but you crossed 100K points landmark today. WooHoo!

[Pierre Lindenbaum]

not really :-) Pierre Lindenbaum 99550

[GenoMax]

Upward rounding of the BioStars code already put you at 100K :-)

[floris.barthel]

I edited several (multi-readgroup) BAM file SM tags this way. After reindexing them, they seem to work fine, ie. I can open them in IGV without errors. However, when I use Picard ValidateSamFile I get the following error:

## HISTOGRAM    java.lang.String
Error Type      Count
ERROR:INVALID_INDEX_FILE_POINTER        1

I thought this wasn't a big deal, because they open fine. However, when I try to use other GATK tools to further process these files I'm often given an error like:

htsjdk.samtools.util.RuntimeIOException: my.bam has invalid uncompressedLength: -1998504602
    at htsjdk.samtools.util.BlockCompressedInputStream.inflateBlock(BlockCompressedInputStream.java:543)
        ...
        at org.broadinstitute.hellbender.Main.main(Main.java:289)

or

htsjdk.samtools.SAMFormatException: Invalid GZIP header
        at htsjdk.samtools.util.BlockGunzipper.unzipBlock(BlockGunzipper.java:121)
        ...
        at org.broadinstitute.hellbender.Main.main(Main.java:289)

Is there anything else I can do to fix this other than regenerate the BAMs from source?

[jtzheng222]

hi! I also have the same problem: htsjdk.samtools.util.RuntimeIOException: my.bam has invalid uncompressedLength

Have you finally solved this problem? and how do you solve it? thanks

[xuyao15927402897]

This is because the .bam.bai does not correspond to the bam file. After you change the @RG tag of BAM, you need to re-build the index, such as:

samtools index xx.bam

good luck!

[floris.barthel]

Any ideas anyone? Pierre Lindenbaum ? Using samtools reheader + sed to change the SM tag for multiple readgroups at once (something that I dont' think is possible with AddOrReplaceReadGroups) is a very elegant solution but if GATK complains about an invalid file it's not a viable one.