I’m using SILVA NR99 138.1 which as far as I can tell holds species level data on sequences. Although I understand the limitations of species-level classification of the 16S gene I’m just curious as to why:
A. A lot of papers state that SILVA and RDP only go down to genus level when their fastas in fact contain species IDs (is this a newer thing from the recent years?) ?
B. I was able to use Illumina’s 16S Metagenomics app to classify to a species level with the RDP database
C. classify.otus results delivered ASVs (Don’t have enough storage for OTUs distance matrix) to a genus level when using SILVA 138.1 containing sequences labeled to a species-level
A. A lot of papers state that SILVA and RDP only go down to genus level when their fastas in fact contain species IDs (is this a newer thing from the recent years?) ?
Their taxonomy database only goes to the genus level. Keep in mind that many genera only have one sequences. I suspect that although individual sequences may have a species name there is often only a single species represented in the genus or each species only has one sequences. To have credible classifications you really need multiple representatives of each species and sibling species.
B. I was able to use Illumina’s 16S Metagenomics app to classify to a species level with the RDP database
I wouldn’t trust much coming out of Illumina’s bioinformatics pipelines.
C. classify.otus results delivered ASVs (Don’t have enough storage for OTUs distance matrix) to a genus level when using SILVA 138.1 containing sequences labeled to a species-level
See A.
Hope this helps some. Regardless of what the tools give you, 16S cannot get you species level information. You need genomic information for that.
Thanks for the advice @pschloss, how would you recommend visualising classification results from classify.otus? As I don’t really trust Illumina’s pipelines either but I’m a novice and don’t really know how to get them nicely displayed in a pie chart or something similar.