"Year" Excluded as Numeric from Kruskal Wallis Test

Hello all,

I am running the following script on my microbiome data:
qiime diversity alpha-group-significance
–i-alpha-diversity dunbar-core-metrics-results/faith_pd_vector.qza
–m-metadata-file /home/mingala/array1/dunbar/dunbar_metadata.txt
–o-visualization dunbar-core-metrics-results/faith-pd-group-significance.qzv

When I view the resulting .qzv file, it tells me that “The following metadata categories have been omitted because the metadata was numerical, not categorical: Year”. I would like to test for differences between sampling years, so how can I tell QIIME to treat this as a factor rather than numeric entry? I see it worked without problem in the Moving Pictures tutorial. Thanks!


Hi @mingala!

Where do you see that at? The alpha-group-signicance plots in the Moving Pictures tutorial exclude the Year category, just as you described above:

We are in the process of overhauling metadata, and we are going to support optional typing of metadata columns, which would allow you to declare the Year as categorical, but that hasn't dropped just yet.

In the meantime, you could trick QIIME 2 into thinking that column is categorical by adding some text to the values (e.g. Y2017 instead of 2017).

Let us know how it goes! :t_rex:


This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.

In the QIIME 2 2018.2 release, Metadata supports optional column typing, allowing users to declare a metadata column “categorical” or “numeric”. Using optional column typing, columns that look “numeric” can be declared to be “categorical”, allowing the columns to be used in categorical analyses such as alpha-group-significance.

There are a number of other changes to QIIME 2 Metadata in the 2018.2 release. See this forum announcement for details on what changed, as well as the updated Metadata tutorial. :sun_with_face: