I was analyzing some 16S sequences with Mothur. After the whole pipeline, I got a big group of Rhizobiales. My PI using the QIIME pipeline got Caulobacterales. For our samples, the Caulobacterales make more sense than rhizobial. It is clear that these are the same sequences. Our hypothesis is that the trainset I am using (trainset9_032012.pds.tax/fasta) classifies the sequences as rhizobial instead of caulobacterales.
Is there any way to use another trainset. How about a SILVA trainset? How can I make a trainset compatible with Mothur? In advance, thanks for your help!
I have stopped using PDS due to problems with my positive controls for taxonomic assignment. I run SILVA.
As a suggestion, you could run multiple taxonomic assignation in Mothur, no problem, and compare them together. A while back, we had better results with PDS for some bacteria and some others it was better with Silva (based on our positive controls).
Thank you for your suggestion. How do I adapt the SILVA’s database to work as a reference? Do I follow the instructions on Mothur (the one that says how to prepare the PDS files for the sequence classification)? Or are there extra steps? Again, thanks!
Hello, I actually do not know. I work with gut bacteria and always use the default files provided by Mothur: Silva reference files (mothur.org)
Best of success
This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.