filtering/extracting/removing commonly/uniquely mapped reads from reference mappings

[yaximik]

Is there a package/program/tool to identify and extract WGS reads mapped uniquely or to both reference (e.g. human and chimp) genomes from resulting indexed bam files?

[MikeChoud]

I dont think so. Probably the easiest approach is to map to one assembly, only report mapped/discard unmapped, convert bam to fastq and then map to the second assembly. As the bam file contains the assembly info in the header I dont see how you would have both.

Now that I think of it, maybe check the workflows for spike-in normalisation. Typically its done with a separate organism so this might give you a solution!

Do tell if you find a solution

[fchatonnet]

Hello,

you may also use simpler approaches with basic text editing tools from UNIX, after converting your bam files into sam (or after a samtools view command). For example, you may make a file with the read names aligned on one reference and use grep -f with this file on the other sam file. You can use option -v to extract the reads which are specific of one reference genome (not found in the other one).

You may also use join on the sam files with several options, using the read names as the joining field (don't forget to sort the sam files). Here also you can use option -v 1 or -v 2 to keep only reads specific from your first (respectively second) alignment.

Good luck!