Samtools Mpileup On Single Bam With Multiple Samples

Question

[Danielk]

I have a fastq file where reads are named after the sample they belong to, like so

@S1-HWI...
AAGATTG
+
qualqual
@S2-HWI...
GGTGAGG
+
qualqual
...

After aligning this data I want to call variants for each sample and would enjoy a VCF file as output. Now to my question:

Is there a way I can use samtools mpileup for the SNP calling? Can I use the RG/LB tags somehow?

This would be extremely convenient in order to handle multiple samples in a single fast file.

cheers

//Daniel

Answer 1

[Pablo]

You should add the @RG header to your SAM file.

Then add the SM tag to each read (which should be easy since it's in the name). A small script like this should do the trick:

#!/usr/bin/perl
while( $l = <> ) { 
  chomp $l; 
  $l ~= /^S(\d+)/ or die "Cannot find SM\n";
  print "$l\t$1\n";
}

Assuming the script is called addSm.pl, you can do:

cat my.sam > ./addSm.pl > my_SM.sam

Here is the quote from the manual:

SAMtools acquires sample information from the SM tag in the @RG header lines.

I hope this helps.

Answer 2

[Swbarnes2]

I don't know if samtools mpileup will split a single .bam into multiple samples, but if you can get each sample into its own .bam, then you can input multiple .bams into mpileup, and it will put all the data side by side. This can be a way to look for SNPs that are common or unique to your samples. Unfortunately, one of the most useful measure, the DP4, is a combined entry. But you can always consult a single sample vcf to get that for a particular sample.