Nematode Identification Through Ngs

Question

[tuomaz84]

Hi,

I'm playing around with sequences I have from amplifying nematodes from fecal samples. What I'm ultimately trying to do, would be identifying nematode-species by their 18S sequences. Now I'm looking for good software to deal with my data. Blast+MEGAN works just fine and it seems my sequences are good, but I'd like to group them to the OTUs - Megan leaves quite a many of them to the level of Nematoda and I'd like to find out if they would be clustering to multiple OTUs.

I've tried PANGEA but apparently that's good only for prokaryotes and RDP (at least I can't modify taxcollector-database from NCBI databases). So, are there any softwares which could be used to pipeline all my samples or do I have to use separate programs to denoise, cluster and identify?

.tuomas

Answer 1

[Dave Lunt]

There are a few ways you could go with this Tuomas. Personally I don't think anything that is BLAST based is a good idea. This is phylogenetic problem at heart. BLAST will be just as good in some situations but make terrible hard to discover errors in others.

QIIME is a very powerful tool, which works with eukaryotes too. I have used this to identify eukaryotic 18S from 454 sequence data. But I'm not sure you need this scale of software. If it comes from PCR amplification and Sanger sequencing you could just align your samples to a reference alignment of nematode 18S and then build a quick tree. I have done this many times using ARB and RAxML and if you email me I'm happy to share the structural alignment etc.

Another approach is to create a reference alignment and tree, and then use either PPlacer or RAxML-EPA to "place" new sequences onto the tree. This is very scalable, and works with a few or lots of samples. These are now both in the latest release of QIIME, I think, though I haven't tried them there.

Best of luck

Dave

Answer 2

[Istvan Albert]

I would also recommend mothur, it is a unique tool in that it is distributed as a completely self contained binary with almost no other dependencies.

Just download and run.