Paired-End Bam Files

Question

[sarahmanderni]

Hi,

Having two BAM files from NGS data, how can one check if they are the BAM files (left and right) from a paired end mapping of the same sample? Thanks for the help.

[Devon Ryan]

The paired-ends from the same sample should (almost*) always be mapped together, producing only 1 BAM file. If you're mapping mates separately, then you're doing things wrong.

*If you were working on an exception to this, you'd already know how to do this.

[ancient_learner]

am not much clear. but if you want to check whether those files are paired-end convert them to sam format using samtools and then see

[lomereiter]

I'm not sure if I understand you right, but maybe sort them by name and check if the read names are the same in the two files?

Answer 1

[Devon Ryan]

Mates in a pair shouldn't be mapped separately. To check if a BAM file has reads that arise from paired-end sequencing (and were mapped accordingly), just check for the 0x1 bit in the FLAG field (i.e., see if samtools view -f 0x1 something.bam prints any reads).

[Chris Cole]

Yup. Agree. If you have separate BAM files for each sample, then either they are separate runs or the paired-end alignment was done wrong.

Have you asked the people who generated the BAMs? They will know what they did, rather than you trying to guess from the files themselves.

[sarahmanderni]

Ok thanks I will ask.

[Chris Fields]

Completely OT, but we really need a way to convert a comment to an answer, this would be the obvious pick for us command-line junkies.

[Devon Ryan]

The option exists for moderators :)

Answer 2

[Shicheng Guo]

I write a perl script to decide the bam file is paired-end or single end.

sub is_bamPE($){
  my($bamfile)=@_;
  my $is_pairedEnd;
  chomp( my $line=`samtools view $bamfile | head -n 1| awk '{print \$2}'`);
  my $remainder=$line%2;
  print "$line\t$remainder\n";
  if($remainder){
  $is_pairedEnd="--paired-end"
  }else{
  $is_pairedEnd="--single-end"
  }
  return $is_pairedEnd;
}

Answer 3

[Chris Cole]

I'd look at the BAM in a graphical viewer. Try one of desktop viewers IGB, IGV or Tablet or online browsers such UCSC or ensembl.

[sarahmanderni]

Actually there are BAM files for more than 70 samples. So I have two bam files for each sample and I want to check if they really are the paired-end files of the samples or are single end but technical replicates. I was wondering of a flag or something that could be checked with a python script to show it.