We now have Silva-based taxonomy classifiers that can be used with q2-feature-classifier's classify method. These are Naive Bayes classifiers trained on the Silva 119 99% OTUs, and we have created these for the 515F/806R region of the 16S and for full-length 16S. These are available from the following links:
- Classifier trained on full-length sequences
- Classifier trained on 515F/806R region of sequences (this file is temporarily unavailable while we rebuild with sklearn 0.18.1)
In the next couple of weeks we’ll be posting documentation that can be used to train these classifiers from other databases, and on different regions of marker gene sequences.
To use these, for example in place of the Greengenes-based classifier in the Moving Pictures tutorial, you would simply provide one of these files as the classifier input to q2-feature-classifier's classify method.
Thanks to Ben Kaehler for training these classifiers!