BAM header fix

[nasepu]

I'm having trouble running a BAM file through GATK. I got this error:

##### ERROR MESSAGE: SAM/BAM file reanalysis.bam is malformed: SAM file doesn't have any read groups defined in the header. The GATK no longer supports SAM files without read groups

So i figure I need to replace the header. I checked the header with samtools view -h, and I see an extremely long list that starts like this:

@HD VN:1.0 SO:unsorted
@PG ID:novoalign PN:novoalign VN:V2.08.02 CL:novoalign -o SAM -d ./reference/chok1ref.nix -f ./cho9/reads.fq
@SQ SN:gi|351517969|ref|NW_003613580.1| AS:chok1ref.nix LN:8779783

and continues with many @SQ lines

I tried fixing the header with Picard using

java -jar picard/AddOrReplaceReadGroups.jar INPUT=cho9/cho9output.bam OUTPUT=cho9/cho9outputH.bam RGID=RSNS002 RGLB=RSNS002 RGPL=illumina RGPU=RSNS002 RGSM=RSNS002

But that didn't seem to generate an output file, and I did just put the same thing in every field there, so I'm a little unsure about it.

So, is it possible to fix the header? If so, what am I missing? If not, do I just need to do the alignment over again and properly set the header using something like:

 bwa samse -f out.sam -r "@RQ\tID:<ID>\tLB:<LIBRARY_NAME>\tSM:<SAMPLE_NAME>\tPL:ILLUMINA" hg19 input1.sai input2.sai input1.fq input2.fq

Help is greatly appreciated!

[westerman]

No help here. You have a nicely complex problem for which there may not be an known fix. Certainly a fix is not obvious to me -- I'd have to play around with the files and you are already doing that at a sophisticated level thus I have nothing to add. But since you are a new poster I did want to say "welcome!". You may not get a response to this question but be assured that the rest of us are listening and will help as we can if you have other questions.

[TiborNagy]

We use the following parameters (version 1.68):
java AddOrReplaceReadGroups.jar I=input.bam O=output.bam SORT_ORDER=coordinate RGPU=na RGID=1 RGLB=input RGPL=Illumina RGSM=Company CREATE_INDEX=True

[vivek_]

You could try adding VALIDATION_STRINGENCY=LENIENT and see if this makes a difference with Picard add or replace read groups.

Do you see any errors or warnings in the stack trace when you run Picard?

[nasepu]

Thanks for the help. Turns out I forgot to give java enough memory for AddOrReplaceReadGroups. Adding -Xmx4g made it run just fine.