Using API interface for q2-sample-classifier

[nricks]

Are there more tutorials on API interface other than this one?
Artifact API — QIIME 2 2018.4.0 documentation
This seems to give a very limited number of options on what you can do with it.
Specifically I would like to use it for the q2-sample-classifier function to predict metadata categories of unknown samples as mentioned here
How to use classify_samples to predict metadata category? - #3 by Nicholas_Bokulich

[colinbrislawn (Colin J Brislawn)]

Hello Nricks,

I don't think there is a tutorial for this, but the documentation does include info about the API as well as the CLI.
https://docs.qiime2.org/2018.4/plugins/available/sample-classifier/classify-samples/

Colin

[thermokarst]

Hey there @nricks!

@nricks

Are there more tutorials on API interface other than this one?

Not yet, but have you seen this community contribution?

Moving Pictures Python API tutorial (QIIME2 2018.4)

I rewrote the "Moving Pictures" tutorial using QIIME2 Python API. The original QIIME2 CLI "Moving Pictures" tutorial can be accessed here. It takes advantage of the latest QIIME2 2018.4 functionalities, especially displaying visualizations in the notebook. You can also view it in nbviewer. This notebook should be treated as experimental and its goal is to provide a starting point for the users in going from the CLI to Python API. t.

@nricks

This seems to give a very limited number of options on what you can do with it.

Our tutorial is certainly very anemic at the moment, but, the general design strategy of QIIME 2 is that it is interface agnostic - interfaces all expose the exact same functionality (which is defined in plugins). That means the DADA2 denoise method is the same when called from the cli interface, the artifact api, or the GUI studio interface!

Let us know if you have any more questions! :qiime2:

[nricks]

Thank you, I've got a much better grasp on how the API system works. What I would like to know, is how exactly to create a classifier using the q2-sample-classifier function, and then feed in a new set of data and have it predict the metadata

[thermokarst]

@nricks

What I would like to know, is how exactly to create a classifier using the q2-sample-classifier function

Have you had a chance to check out the tutorial?

This plugin doesn't work the same as the feature classifier plugin at present - training and evaluation occur in the same step. We have an open issue to support training classifiers independently, but not sure on an ETA for this feature. Stay tuned!