Gian77
commented
1 year ago Hello @szymanskishay,
It is kind of hard to help withough knowing which versions of CONSTAX2 and UNITE are you using. Please run constax -v to check which version you are using. Additionally, how do you run constax?
Regarding your first question, Iw ould try rerunning constax using the --train option again and see what happend.
For the second question, @liberjul may have an answer for you.
Thanks for using CONSTAX2! G.
szymanskishay
In running a CONSTAX job on the MSU HPCC there were some strange results in taxonomy assignment. For details I do have to give that I suspect may be relevant, I had the job train using the UNITE general fasta releases for all Eukaryotes ITS. Specifically, I am running into a situation where there are spurious _1's throughout certain OTUs (Ex. Basidiomycota Agaricomycetes_1 Agaricales_1 Cortinariaceae_1 in one OTU, but another that has these same levels without the _1) which I assume is a holdover from the appending things for fixing duplicates to not break RDP classifier. However, these things weren't the type of duplicates detected as an issue by the classifier.
Additionally, many OTUs have their assignments at the wrong taxonomic level. For example, I have some OTUs that belong to the "Kingdom" Alternaria metachromatica. One instance is Kingdom: Basidiomycota Phylum: Agaricomycetes_1, Class: Agaricales_1. Unfortunately, I don't have the slurm output for this file anymore to attach for help, but I was moreso curious if this type of error had been seen before. To me, it looks like occassionally lower taxonomic levels are being considered Kingdom and everything else is in the proper descending order but being shifted over a few slots (such as family becoming Class, class becoming Phylum, and Phylum becoming Kingdom in the above example).