mapping of metagenomic data to multiple reference genomes

[hsberro]

Hi,

I am trying to map metagenomic data to a set of more than a thousand genomes.
I expect many of the reads not to align to anything as it is metagenomic data and there are many bugs out there that do not have a reference genome.

I am not sure what would be the best alignment tool in this case. Was thinking about Bowtie2 but as far as I understand I will need to index each of the reference genomes separately and run each of my samples against each of the indexes.

Any suggestions?

Thanks,
Marie

[Brian Bushnell]

This depends on your specific goal. You can always just concatenate all the references together in a single file, and then index that, which would allow you to use any mapper, with a single pass. You'd have to ensure all of the scaffolds have unique names, though.

Alternately, I have a program called BBSplit designed for mapping against multiple references simultaneously, for the purpose of binning reads - it outputs one file per reference, containing all of the reads that mapped best to that particular reference; in addition, it generates statistics about the fraction of reads that aligned to each, so you can estimate the composition of the community.