How to train SILVA 128 feature classifiers

[gregcaporaso]

Please see the following post for details about training SILVA 128 feature classifiers:

Processing, filtering, and evaluating the SILVA database (and other reference sequence data) with RESCRIPt

Please consider this tutorial a living document, which may change based upon community feedback and ongoing plugin development. RESCRIPt RESCRIPt (REference Sequence annotation and CuRatIon Pipeline) is a python package and QIIME 2 plugin for formatting, managing, and manipulating sequence reference databases. This package was designed for compiling, manipulating, and evaluating sequence reference databases from SILVA, NCBI, Greengenes, GTDB, and other sources, and for construc…

Original Post Content (post links are no longer valid)

I created Silva 128 99% OTU classifiers trained on full length sequences and on the 515F/806R region only. These are derived from the QIIME-compatible SILVA release for SILVA 128.

You can find these files in Dropbox for now, and we'll link these (or more recent classifiers if they're available) from the data resources page for the 2018.4 release.

The most relevant files are:

• silva-128-99-nb-classifier.qza : trained on full-length 99% OTU representative sequences, and the SILVA_128_QIIME_release/taxonomy/taxonomy_all/99/taxonomy_all_levels.txt taxonomy file
• silva-128-99-515-806-nb-classifier.qza : trained on 99% OTU representative sequences, trimmed to the region bound by the 515/806 primers, , and the SILVA_128_QIIME_release/taxonomy/taxonomy_all/99/taxonomy_all_levels.txt taxonomy file

Those are prepared in the same way that we prepared classifiers for the SILVA 119 release.

For experimental purposes, I also trained these classifiers on the the SILVA_128_QIIME_release/taxonomy/taxonomy_all/99/majority-taxonomy-7-levels.txt and the SILVA_128_QIIME_release/taxonomy/taxonomy_all/99/consensus-taxonomy-7-levels.txt files, to experiment with these. These are files that collapse the full SILVA taxonomy to exactly seven levels (domain, phylum, class, order, family, genus, and species), so may be easier to interpret than the others. I'm interested to hear whether folks think these files are useful.

[mcz (maochengz)]

hi, SILVA 128 classifiers can’t be downloaded because the file is missing and I need it, thank you.

[thermokarst]

Hi @mcz - I have edited the original post with the current recommendation on getting SILVA 128 - please see this post for more details:

Processing, filtering, and evaluating the SILVA database (and other reference sequence data) with RESCRIPt

Please consider this tutorial a living document, which may change based upon community feedback and ongoing plugin development. RESCRIPt RESCRIPt (REference Sequence annotation and CuRatIon Pipeline) is a python package and QIIME 2 plugin for formatting, managing, and manipulating sequence reference databases. This package was designed for compiling, manipulating, and evaluating sequence reference databases from SILVA, NCBI, Greengenes, GTDB, and other sources, and for constru…