mapping with botwie/bwa

[Francy87]

Hello!
I'm sorry for the stupid question, but I'm new in metagenomic and metatranscriptomic stuffs... I'd like to map my Hiseq shotgun metatranscriptomic reads to some reference genomes (I already know the microbial community composition, since I analysed 16S data), using botwie or bwa. The question is: do I have to map each sample to each genome? or is there a way to map to more genomes at the same time? And if not, how can I merge the multiple mapping files from the same sample, then?

Thanks a bunch
Francesca

[jimmybee]

Hello!
I'm sorry for the stupid question, but I'm new in metagenomic and metatranscriptomic stuffs... I'd like to map my Hiseq shotgun metatranscriptomic reads to some reference genomes (I already know the microbial community composition, since I analysed 16S data), using botwie or bwa. The question is: do I have to map each sample to each genome? or is there a way to map to more genomes at the same time? And if not, how can I merge the multiple mapping files from the same sample, then?

Thanks a bunch
Francesca

No you wont have to map them separately. Create a multi-fasta file of all your reference sequences and then build the mapping index from that.

[Francy87]

No you wont have to map them separately. Create a multi-fasta file of all your reference sequences and then build the mapping index from that.

Hi Jimmy,
Thanks for your reply. Can I just 'cat' all the reference genomes together or should I modify them in some way?
Thanks again
Francesca

[swbarnes2]

Cat them together, then you use bowtie-build or bwa index to make the index. Then you align.

[jimmybee]

Hi Jimmy,
Thanks for your reply. Can I just 'cat' all the reference genomes together or should I modify them in some way?
Thanks again
Francesca

Yeah no need to modify them.

[Francy87]

Hi again!
I downloaded some reference genomes from the ncbi ftp site (about 500). Is there a simple way to cat all of them recursevely without doing it one at the time? I have many directories, each one with one genome in .fna format...

Thanks

[dpryan]

If all of the *.fna files are only buried one directory deep, then something like the following bash script should work:

#!/bin/bash
dirs=`find . -mindepth 1 -maxdepth 1 -type d`
for d in $dirs
do
    cat $d/*.fna >> genomes.fna
done

This could be modified to easily handle the presence of subdirectories.