Biostar Beta. Not for public use.
BAM file comparison
1
Entering edit mode
2.6 years ago

Hello Biostarers. I am trying to compare BAM files for similarity. I am doing this to assess alignment from different pipelines. Is there a quick straightforward way to do this?

alignment BAM • 1.2k views
1
Entering edit mode

What exactly are you trying to compare in them? Do you want to compare each alignment or just come up with a correlation score?

0
Entering edit mode

I want to assess how similar the alignments are from files for the same sample but from different pipelines. An overall correlation score would be enough I guess.

4
Entering edit mode
16 months ago
Botond Sipos ♦ 1.7k
United Kingdom

The wub package has a script for comparing BAM files. It does a coarse comparison of alignments (checks whether the start position of the alignments is close enough) and also calculates a "consistency score" (the fraction of bases which align the same way).

The package also has a simple long read simulator, which can give you the "true" alignments in SAM format.

3
Entering edit mode
20 months ago
Hussain Ather • 920
National Institutes of Health, Bethesda…

You could try using bamcoverage from deeptools to convert them to bigwig files and then visualizing them using IGV or UCSC Genome Browser.

2
Entering edit mode
15 months ago
Freiburg, Germany

From BAM files, the simplest route would be multiBamSummary followed by plotCorrelation from deepTools. If you want a bit finer grained control, then you can use bamCoverage and the multiBigwigSummary before plotCorrelation instead. Full disclosure: I'm one of the deepTools authors.

0
Entering edit mode

I'm looking to do something similar, but instead comparing bam-files of the same input reads mapped to different assemblies to see the fraction of reads mapping to all assemblies. I was thinking to pull out the names of all mapped reads from each bam-file and then maybe simply make a venn-diagram of the lists, but could I use deepTools for this perhaps?