Not sure which questions to be asking, or how to answer them (metagenome, metabolism)

[grem]

I've been tasked with analysis of a metagenome sample from an extreme environment. I'm a bit new to the meta- stuff, so I opted to initially run the MetAMOS pipeline, which has assembled and spit out taxonomic information.

While this is a good starting point, I want to dig deeper and look at the uniqueness of some enviro-specific things (like maybe sulfur metabolism pathways?), but I'm not sure which way to go or how to get there.

I've done a CDSearch and genefinding via prodigal off of the assembled contigs (I think I'm headed in the right direction here) but I'm not sure what to do with the output.

Any input/ideas?

[mikesh]

As far as I know MetAMOS also finds ORFs and annotates them using BLAST. After that you should get GO terms for matched proteins and do your GO enrichment tests.

[grem]

So using something like this? http://go.princeton.edu/cgi-bin/GOTermFinder

I wonder...does it matter that these types of tools assume a single organism? Are there any ontology tools specifically geared towards meta samples.

Thanks for the help!

[mikesh]

So using something like this? http://go.princeton.edu/cgi-bin/GOTermFinder

I wonder...does it matter that these types of tools assume a single organism? Are there any ontology tools specifically geared towards meta samples.

Thanks for the help!

Actually there are many tools for gene ontology enrichment analysis. You can also simply use hypergeometric tests with some multiple testing correction (eg Bonferrony), or check out EASE framework (http://david.abcc.ncifcrf.gov/).

As for single/multiple organism, I believe it is quite strange to check if proteins in some organism are enriched for some metabolism (e.g. Rhodobacter could easily switch among different types). One better check for proteins of some specific metabolism are enriched in bulk proteome, e.g. like done here http://www.biomedcentral.com/1753-6561/5/S2/S9. Anyways there are lots of possible variations of statistical tests that could be performed here and I dont think some gold standard exists..

[bastianwur]

Why only go for GO terms?
Just throw everything into InterproScan (case you have the computing power), get the EC numbers, and check what the pathway databases say.

EDIT: oops, sorry for the necro *ahem*.