qiime longitudinal linear-mixed-effects --p-metric issue

Technical Support
1

Dear developers,

I am running QIIME2-2019.10 with conda and I am trying to lunch qiime longitudinal linear-mixed-effects on q2-picrust output tables. I've 3 output tables ec_metagenome, ko_metagenome and pathway_abundance obtained with q2-picrust.

I've trouble with the --p-metric names as in the case of ec_metagenome, features are in this format EC:4.2.1.32 and and in the case of pathway_abundance, features are like FAO-PWY. I've tried to write them between " " and between ' ' but
with ec_metagenome response is:

Plugin error from longitudinal:
metric must be a valid metadata or feature table column.

with pathway_abundance response is:
Plugin error from longitudinal:
Error evaluating factor: NameError: name 'FAO' is not defined

It seems that symbols - (minus), : (colon) and/or . (dot) are causing the problem because with the ko_metagenome table where features are like K05311 everything goes well.

Thank you for your support

2

Dear all,
I've modified symbol "-" with "_" and everything goes fine. Just let you know. For readers:
1 I've extracted the feature-table.biom from the qza archive (it is inside folder "data");
2 I converted to tsv (biom convert -i feature-table.biom -o file.tsv --to-tsv);
3 Modify symbol with gedit;
4 convert tsv to biom (biom convert -o feature-table.biom -i file.tsv --to-hdf5);
5 Substitute old feature-table.biom with the new feature-table.biom inside qza
6 run plugin

Cheers

3

Hi Carlo,

That's right — some special characters cause errors because they are interpreted as mathematical symbols in your formula!

no need to extract though. A couple solutions:

  1. (maybe) quoting the metric might work (this should work with the Artifact API and maybe quoting twice would work on the CLI, e.g., "'metric'")
  2. (definitely works) use qiime feature-table group to rename your metric(s) in the feature table, instead of extracting to manually rename.

Here's how to relabel (just apply to the feature axis instead of sample axis):