search

biocore / songbird

Vanilla regression methods for microbiome differential abundance analysis
BSD 3-Clause "New" or "Revised" License
58 stars 25 forks source link

include colinearity discussion? #29

Open lisa55asil opened 5 years ago

lisa55asil commented 5 years ago

should we flag new users about problems with possible colinearity issues if their metadata categories are too similar (e.g. including a continuous value of BMI and a quartile category of BMI)?

mortonjt commented 5 years ago

Not exactly sure how to do that in the context of multinomial regression yet.

Maybe we could just do a tutorial with biplots / summary statistics. There is now a pseudo R2 statistic in the qiime2 commands that allow for those sorts of investigations.