Using qiime taxa collapse Before Logistic Regression: Is It Appropriate?

[jwdebelius (Justine Debelius)]

Hi @jau,

I think the answer is It Depends on several things.

@jau

I wanted to ask if it would be appropriate to use qiime taxa collapse before using the data for a logistic regression

The answer to this is yes. You can absloutely collapse your data to genus level before doing... whatever. You're making a set of assumptions which are discussed in the linked posts below (these are are a starting place.)

Multiple entries for the Same Taxa and How do I retrive the corresponding Fasta sequence for each taxa

Dear All, I have two questions. I downloaded csv file from taxonomy.qzv visualization. I got the feature ID , taxon and confidence. Here i Have a doubt. When I check for the taxa information, I found that same taxa presents more than one times with different feature ID and different confidence. Why is it so? I downloaded .csv from taxa-bar-plots.qzv. How to get the fasta sequences of corresponding taxa? [Screenshot%20from%202018-07-04%2010-49-55]

How should different ASVs of the same species be interpreted?

Hi guys! I understand this is a common question and discussed numerous times on different forums as to why this phenomenon happens and how to solve it (ignore it), but it would be nice to hear from some of you researchers as to how you interpret them in your research? Data coming out from ASVs methods such as dada2 is common to have multiple ASVs classified to the same species, most likely dominant species in the microbiome. My question is, is it generally accepted to treat these ASVs as diff…

There's a second issue in your question, and that's about the appropriate use of logistic regression in microbiome data. I assume you 're using the taxa to predict an outcome? This might also be something to consider. I appreciate that the standard consideration in models that if your exposure (e.g. microbiome) should predict your outcome (e.g. disease) but this can be more complicated.

If helpful, there's also a discussion about that on here:

https://forum.qiime2.org/t/modeling-bacterial-features-as-indepedent-variables/25766/8

Again, happy to discuss further in general terms and as always happy to have others involved!

Best,
Justine