From TheSeed
Revision as of 11:46, 20 June 2008 by FolkerMeyer (talk | contribs)
Jump to: navigation, search

What are projects?

Projects are related sets of metagenomes. If you for example plan on studying a set of samples from a chronoseries, it might me useful to group them into a project.

What level of privacy does MG-RAST v2.0 provide?

We provide password control and the ability for the submitting entity to control access on a username/password basis to the submitted data sets.

Do you support BLASTing against my private database XYZ?

We currently do not explicitly support this, however the underlying software design end system architecture support this.

How frequently do you update the underlying NR for MG-RAST?

With version 2.0 we have added the support for multiple concurrent sets of sequence similarity results to be stored per metagenome. We can add results for newer NRs. However once you start comparing results for metagenomes (say you are interested in the phylogenetic reconstruction) different versions of the NR used for the underlying data will lead to incorrect comparison results as older versions of NR will miss certain organisms and or annotations.

How long does it take to analyze my Metagenome?

The answer depends on two factors a) the size of your data set and b) the current server load. Under optimal conditions, it takes about 18 hours to run a 100 million basepair 454 metagenome through the pipeline.

What parameters should I use to analyze my data?

The answer depends on your sample. In any case we recommend that you modify e-value, minimal alignment length and percent identity requirements for the BLAST results underlying the results. The effects of this are different for each sample. Depending on sample complexity, sample size, number of species an diversity of species present your results will vary dramatically when modifying these parameters. For RNA based phylogenetic reconstruction, we recommend requiring a minimum alignment length of 50bp for exact matches.