Filtering biplot for taxa

[josmos (Josef Moser)]

I was creating a biplot with more than 4000 features, which is convoluting my plot.
Is there a way to filter them in order to show only the important ones?

[jwdebelius (Justine Debelius)]

Hi @josmos,

You can try a couple of things!

You could try passing a table collapsed to a higher level: phylum or class often limits your resolution.

If you've run ANCOM or some other feature-based test, you could select the significantly different features, and then plot those.

Best,
Justine

[josmos (Josef Moser)]

So i filter the --i-features for the taxa i am interested in, right?

[jwdebelius (Justine Debelius)]

Hi @josmos,

Umm... as I'm looking at the qiime 2 2019.1 release, I'm not seeing a --i-features flag for qiime emperor biplot. Were you referring to this command, or to the filtering step?

Best,
Justine

[josmos (Josef Moser)]

Sorry, I meant the qiime diversity pcoa-biplot.

[jwdebelius (Justine Debelius)]

@josmos,

Thanks for the right command! You'll want to pass your filtered feature table into the --i-features flag.

You can get to that filtered table via the qiime feature-table filter-features function, and just pick your list of interest either from ANCOM or most abundant, or whatever else makes sense for you. (I'm going to recommend ANCOM again because I think its most likely to give you the information you're looking for).

Best,
Justine