Marcio, You are right there is something a little bit odd about these jobs.

From the website I do see that that the protein search came up with about just as many hits as the rna search ( http://metagenomics.anl.gov/metagenomics.cgi?page=MetagenomeOverview&metagen...). Is this amplicon or shotgun metagenomic data?

The other odd thing is later run through the system (4454965.3) only 2104 sequences passed the quality filtering. Which actually lines up nicely with the combine number of hits found between protein and rna. Potentially this could be a sequence quality issue? I'm having trouble logging into our system from home tonight so I can't get at any more details than that but I will take closer look tomorrow. A side note, in the future if you could address emails to [email protected] we would appreciate it. This will make sure you get an answer quickly in case one of us it on vacation or unavailable. Cheers, Jared On Tue, Apr 5, 2011 at 5:00 PM, Marcio Costa <[email protected]> wrote:

Dear Jared,

I have a question about a metagenomic study of the bacterial 16S r-RNA that we had uploaded (4454965.3) on MG-RAST. This job just became available to me on the new version, however, despite been showed that it had 17,593 sequences analyzed on the browse metagenomes page, it gives only around 1000 reads on the metagenome analysis page. In addition, when I compare the results with the previous version of MG-RAST, they are completely different, even using the same dataset (RDP). This same file has been uploaded in the new version (4459773.3), but even less sequences are showed (around 600).

The same issue has happened with several other files we had previously loaded.

Please let me know your thoughts.

Thank you,

Marcio.