Sequence classification in MOTHUR vs QIIME

I have tried to classify some sequences from some .fna files with very short leght sequences (70-80 bp usually) using the classify.seqs script. On top of the poor classification when I go deeper in taxonomic level, I get very different results when I classify against different db (GG or SILVA), and within the same db, usign different methdos, such as de Wang or knn.

I have also used qiime to do the same task and different methods seem to converge better (being able to classify deeper, with more diversity and with compatible results for different approaches).

I am not sure about what can be the reason for all of that.
Thank you.

I hope you can appreciate that if you use different databases and different techniques, you will certainly get very different methods. How about approaching it scientifically and change one variable at a time? Once you have a more specific question, let us know and we’ll be more than happy to help figure things out.