'qiime longitudinal volatility' needs clarification for --p-metric

I am looking at various alpha diversity metrics using qiime longitudinal volatility. The example used is shannon diversity and the metric to use is obviously shannon. But when it comes to using the evenness metric, it is unclear that the --p-metric should actually be pielou_e which is not intuitive. I figured this out by using qiime tools extract to check the title of the vector. This should be clarified in the --help docs and/or tutorial. It would also be nice to have some beta diversity examples. I was having trouble with that (possibly confusion with the metric as well) although I haven’t looked too much into it. That’s just a ‘nice to have’ I figured I’d mention too. Thank you!

Edit: to expand, I used qiime diversity core-metrics-phylogenetic so the evenness vector is a standard output. If I created the evenness vector on my own, I might have seen the ‘pielou_e’ metric but I’m guessing most users will produce the core-metrics.

1 Like

Hi @Jennifer_Fouquier,
Thank you for the feedback! I definitely agree with you here, valid metric names can be a little hidden, especially in cases such as evenness, as you describe. I am working on better ways to describe or auto-detect this information.

For now, the best way to figure out valid metric names (other than cracking open your QZAs to see what’s inside) is to use metadata tabulate.

I have an open issue to provide better description in the docs tutorial (which in the short term will probably just be the advice I offer above).

Thank you for raising this issue!

2 Likes

The QIIME 2 2017.11 release has an expanded description of --p-metric for q2-longitudinal volatility analysis! :tada:

1 Like