I’ve had a couple 16S-V4 experiments and used mothur successfully for the analysis. Now, we’ve shotgun sequenced and assembled genomes of specific isolates of interest from a “sister experiment”. What I’d like to see now is if the 16S sequences of the genome assemblies of these isolates will cluster together with some of the OTUs I’ve obtained or form separate OTUs. Asses the relatedness with a tree then.
Any idea what the best way would be to go about this with my existing mothur pipeline and results?
Here’s approaches I could think of, from stupidest to not-so-stupid I guess:
- prepare 16S-V4 libraries from each of these genomes, repeat mothur analysis with these additional samples
- simulate error-free 16S-V4 reads from each of these genomes, repeat mothur analysis with these additional samples
- pull out the representative 16S-V4 sequences from mothur OTUs of interest (same taxon as genome assemblies), add the genomic 16S-V4 sequences, reallign to the database and recluster independently with opticlust? I would split each mothur OTU from the previous analysis and treat it as a single sample for the new analysis to avoid dealing with different nr of seqs across biological samples.
I have an idea how to do 3) with UPARSE, but taking opticlust OTUs and then feeding them into a different clustering algorithm sounds bad.
My main problem for 3) is that I’m not sure how to get to the cluster.split step without supplying all the taxonomy and preprocessed FASTA sequences. To answer my question I don’t, taxonomy is not important and I would just need to allign the sequences against each other and not a database. Not sure if this is compatible with the existing mothur pipeline.
Any hints/pointers? Thanks.