GATK and SAM format

[dkrtndhkd]

I want to call SNP/indel using GATK tools.

however, GATK occurs error because my input bam file is malformed

because some of @HD, @SQ, @RG, @PG info are missing.

I got the meaning of each component from 'the sam format specification(v1.4-r985)' pdf file from samtools,

but i don't know exactly what should i type each information and wht's critical for GATK because there are many options to fill it.

does anyone who knows it? is there any website to specify it automatically?

[sdvie]

Hi!

From which aligner do the bam files come from?

I have never tried this but in samtools there is the option reheader, where you can replace (or add?) the header.
You would have to prepare a file containing the header lines based on the information that is specific for your alignment = sam/bam file.
You would have to include @HD (including the sort order of your alignments), @SQ (information on the reference sequence) and @PG (information on the program used for alignment).
For the @RG lines, I think you could user the AddOrReplaceReadgroups command from picard.

maybe that helps...
cheers!

[dkrtndhkd]

I use both bwa and novoalign. i use data of illumina Hiseq, pair-end.

Thank you for your reply, Sdvie.

and if you don't mind, may i ask you a example sam file you had used without error?

i do want the header section only, not aligned information.

if you don't want to show others, please send me an email [email protected] ...

Thank you

[sdvie]

I sent you an example of a header. You can see that the SQ line is multiplied because the reference sequence fasta and index has these subsections by chromosome etc. It is crucial that they coincide with the reference sequence you are using in all steps (alignment, GATK etc). Therefore, usually the aligner program itself adds these lines, I have never done this manually. bwa for sure does add the HD, SQ and PG headers, I have not worked with novoalign. As I said before, the read groups may be added either via a bwa argument (check the bwa documentation) during alignment or by picard afterwards.