Gneiss use of combined categories in covariates


for my master thesis I am trying to find a correlation between different feeding strategies and the microbiome composition of different animals.
On the one hand I have samples which were fed with just one feed (e.g. beef), and other samples got a combined feed (e.g. beef AND poultry). I have two different solution approaches in the attached mapping file (column H and columns I-L). What is the best strategy telling gneiss to look for combined categories as well as the effect of each category separately?

mappingfile_feeding_qiime2.csv (7.9 KB)

Thank you very much!

1 Like

You can specify multiple covariates under the --formula option under the ols-regression command. If you want to look at beef and poultry simultaneously, you can specify the formula as follows
--formula C(cow) + C(poultry)

There are many different ways to mix and match covariates. I would refer to the patsy docs on more details (since we are utilizing their statistical formula framework :

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