filtering out human contaminants

[ruchi]

Hi,
I am an experimental biologist and is very very naive to bioinformatics. I have been trying to analyse metagenome data from one of our sample. We are interested in understanding the gene pathways represented up or down in relative samples. We are following the workflow which was done in HMP. To filter out human contaminants i used bmtagger but i got the file output of the contaminants.
Does anyone know how to extract the unaligned files after running bmtagger?
Thanks
Ruchi

[Brian Bushnell]

I have never heard of bmtagger, but at JGI we use this pipeline for human contaminant removal. It's quite thorough, and carefully designed to have zero false-positive removals when processing any domain other than animal. The output is two fastq files, one with the human reads and one with the clean reads, and it keeps pairs together.

[ruchi]

Thanks Brian for your reply. BMTagger was mentioned on the human genome project site so i as trying it out. But no I will try BBMap and if encounter any problem, will let you know.

[MicroBio]

We use DeconSeq for human contamination removal. It has an online application as well as a stand-alone version so you can try it out online before spending the time installing it. Unfortunately it looks like their servers are having troubles so you might have a hard time getting to use the online application.

It outputs a fastq with your clean sequences and one with the contaminated seqs. It's definitely worth checking out.