Since we released version 3 we have had to made a few changes address bugs but the�fundamental�annotations and the counts should be relatively unchanged. One important thing to note with the analysis in v3 is that sequences are mapped to proteins that can occur in multiple organisms. Unlike in version 2 there is not a 1 to 1 relationship between sequence and�organism�abundance. This is especially the case with proteins from core metabolism.�Additionally�the table on the analysis page when it groups by a taxa level it�aggregates�all the counts from lower taxa levels and so unless you have the table set the same way the abundances can appear different.�

I'm not sure if this is related to what you were seeing but currently there is a bug if the annotation source set to M5NR. It will include all of the protein sources as individual rows on when the table grouping is cleared but�aggravate�the counts together on grouping. We are fixing this with an update that will go live later today.�

In the case of the rna annotations (SSU, RDP, Greengenes)�we dropped sequences without�taxonomic�annotations from our rna database because we found the data sources have included large numbers of un-annotationed�environmental sequences. The result is that you should find more of the rna hits having a full taxa annotation but the counts should remain relatively the same.�

I hope this addresses your questions. If something is still unclear please do no�hesitate�to ask.�

Cheers,

Jared Wilkening

- for the MG-RAST development team

On Tue, Apr 5, 2011 at 10:10 AM, <Gomez-Alvarez.Vicente@epamail.epa.gov> wrote:

Dear MG-RAST;

Hi, back in March 24 I download the results of the "Organism Classification" (class and species) analysis for the following projects 4459783.3, 4459784.3, 4459792.3 and 4459793.3. �I used the SSU, RDP and SEED annotation sources. �Today I download again the results/analysis using the same parameters as before and for my surprise the number of sequences annotated �as organisms were different and in some cases decrease by more than half from the original analysis of 03-24-2011.

Do they reanalyzed the Organisms Annotation? �Or change the parameters for annotation? �Since the Functional Classification (Subsystem, COG, and KO) do not suffer any change that I notice.

If any time I redo the analysis it will show a different results? �At least for the Organisms Annotation?

Thanks again for this service

Vicente Gomez-Alvarez
Federal Post-Doctoral fellow
Microbial Contaminants Control Branch
Ph: 513-569-7362
Room: 302

Mail:
U.S. Environmental Protection Agency - ORD
National Risk Management Research Laboratory
Water Supply and Water Resources Division
26 W. Martin Luther King Dr. [Mail Stop - MLK 387]
Cincinnati, Ohio 45268