How to identify every read in a fastq file?

Question

[gerberd1990]

Hi, Everybody is looking for their target reads in a fastq file, and I am just sitting here and can not find a good program to identify the remaining (junk) reads. I am working on ancient DNA (currently horse) illumina reads, and I want to identify the exact organisms (possibly pathogens, human or other contamination, etc) of the remaining reads besides the horse sequences (approx 20-30% of the data contains horse DNA actually). So, can anyone recommend a good program for this task? Thanks in advance :)

[gerberd1990]

WOW, thanks for everybody, I see many valuable information here :)

Answer 1

[Medhat]

First you need to align your reads to the expected reference that it may be contaminated with using FastQ Screen Or DeconSeq, then below post to remove it

http://seqanswers.com/forums/showpost.php?p=109308&postcount=6

Answer 2

[Simon Cockell]

This would seem like a good use case for something like Kraken: https://ccb.jhu.edu/software/kraken/

But your ability to assign every read in your experiment to an organism of origin will depend entirely on the completeness of your database.

Answer 3

[WouterDeCoster]

This might be worth trying: https://github.com/smangul1/rop, see paper: http://biorxiv.org/content/early/2016/05/13/053041

[Medhat]

I think it is specialized in:

discover the source of all reads, which originate from complex RNA molecules, recombinant antibodies and microbial communities.

[WouterDeCoster]

Yes, it is not meant for this application but might give some ideas.

Answer 4

[igor]

There are some suggestions in these previous threads:

• How would you discover the source of contaminant DNA ?
• Faster BLAST alternative