Hi there! I have two quick question: 1. Is there c...
# ampliseq
e
Hi there! I have two quick question: 1. Is there currently a way to call ASV/OTUs for full length 16S amplicons from Nanopore sequencing (SUP Basecalled with latest chemistry)? 2. For epa-ng phylognetic placement is there maybe a resource with some reference trees I could use out of the box (eg a tree and msa from the SILVA DB)? Thanks!
d
1. I think there was recently a new method mentioned in the channel. 2. For phylogenetic placement I am providing reference trees and alignments for the GTDB trees here: https://figshare.scilifelab.se/articles/dataset/SBDI_Sativa_curated_16S_GTDB_database/14869077. The reference sequences were vetted with Sativa to make sure they are consistent with their taxonomic assignment. After this, one sequence per species was selected, aligned and the phylogenomic trees from GTDB were reoptimized with the alignment (see description in the Figshare entry). We also provide sets of vetted sequences meant to be used by other taxonomy assignment algorithms available through the pipeline.
#CEA7TBJGJ is supposed to be able to place ASV sequences in the GTDB trees automatically, but there's a bug I'm currently working on; will be fixed in the next release. For standalone placement there's the #C04NLA7QHH9 pipeline that uses epa-ng.
d
About 1.: not yet within the pipeline, at least not officially supported. There is a promising tool for ONT ASV generation in the development though, as discussed in https://github.com/nf-core/ampliseq/issues/741, but its a bit too early to add it to ampliseq, maybe later this year.
e
Thank you both! The ASV Tool looks promising
I used the SBDI Tree with epa-ng and worked really well thank you!! Was exactly what I was looking for! Do you have any recommendations on how to manage memory usage better? epa-ng used peak ~ 85GB of memory for your tree. @Daniel Lundin
d
Unfortunately, I don't think there's much you can do about that. There is close to 190,000 bacterial species in GTDB (not all are included in my tree since we don't have 16S sequences for all) and if you want to place sequences in that tree it will take a lot of memory. Since a while, GTDB-Tk has split its tree into multiple subtrees -- one per phylum or so -- to ease memory requirements. That would be a possibility, but there's quite a bit of work involved in that.
🙌 1