Missing Methanomassiliicoccus

[sam....@gmail.com]

Hi Falk,

I have run into a misclassification problem with my archaeal data set. I did the analysis with  LotuS 1.502 using the default values for classification i.e. I did not pass any of the options for taxonomy. I got a number of OTUs classified as “Thermogymnomonas” and “unclassified Thermoplasmatales” sequences. I know these shouldn't exist in the environment I sampled and Methanomassiliicoccus should. Both are in the same class but are different orders. If I take the representative sequences from these OTUs and assign taxonomy through the RDP website or I reassign the taxonomy using QIIME they are definitely Methanomassiliicoccus. My question is what is happening when you use the default values for classification? The log file indicates RDP is being used but this gives a different result to the RDP website.

I also have a question about the biom files. I always like to convert the biom file to a txt file so I can have a look at it using the biom convert command (example below) but this never works properly with biom file from lotus. Are they in a different format from other biom files?

biom convert -i OTU.biom -o classic_otu_table.txt --header-key taxonomy --table-type "OTU table" --to-tsv

Thank you

Sam

[Falk Hildebrand]

Hey Sam,

sorry for the late answer, just saw that my email alerts were deactivated for this group. If this taxonomic group was not in the RDP training set, then it could be that it is not being assigned. If you look in the lotus output dir, you will see the actual raw RDP output, as it is run locally on your harddrive. I suspect that the website had newer RDP version, as they only release infrequently also updated databases with their locally executable classifier. However, in my experience only for rare groups this could be a problem. 

Did you try to assign the taxonomy using lambda similarity assignments to SILVA?

About the biom format: I normally use tsv files as primary output in LotuS, you can find them in "higherLvl" dir in the output directory, or directly the OTU.txt file in the outputdirectory. I also updated the .biom format recently, maybe by now it loads the table correctly?

hth,

Falk