Understanding Gneiss Outputs

[Fabs (Fabiola)]

Hi,

I have a quick question. I am trying to determine what taxa are significantly different per trtm, Units, etc. I used Gneiss, and I was looking at the outputs, but I am a little confused. Are the taxa that are shown, y0_taxa_summary5.qzv (358.3 KB)
technically the differential taxa within the variables chosen? Does this also mean they are also significant, if not, what test can I run here, or do I need to use R?

Thanks

[mortonjt (Jamie Morton)]

@Fabs, I’m not completely following what you mean per trtm.

The way that this works is that gneiss will test differences between pairs of groups of taxa via balances. The significance is evaluated in the regression summaries generated in a previous step.

Did you have a chance to look over the gneiss tutorial?
https://docs.qiime2.org/2018.6/tutorials/gneiss/

If there are still confusing details, maybe we should address them there.

[Fabs (Fabiola)]

I did look over the tutorial, but I am still confused. I understand that it looks at the difference between groups of taxa, but is it also differences between groups taxa based on the variable we input. Per say, I did the regression to look at “Time Since Fire” and it came out as 0.05 p value so not significant, and when I did the balance taxonomy, I used the “Site Column” and those were the results I send you. Hence, my question is, do the taxa that are shown here, are they then the taxa that are explaining the difference per site?

Per the tutorial:
“Specifically we’ll plot a boxplot and identify taxa that could be explaining the differences between the control and patient groups”

Sorry, I am really confused in understanding the results.

[mortonjt (Jamie Morton)]

@Fabs

I did look over the tutorial, but I am still confused. I understand that it looks at the difference between groups of taxa, but is it also differences between groups taxa based on the variable we input. Per say, I did the regression to look at "Time Since Fire" and it came out as 0.05 p value so not significant, and when I did the balance taxonomy, I used the "Site Column" and those were the results I send you. Hence, my question is, do the taxa that are shown here, are they then the taxa that are explaining the difference per site?

Yes - the taxa in the numerator and the denominator of that ratio the ones explaining the differences in that category.

Does that answer your question? Let us know if you need more information.

I also recommend reading this paper to clarify additional confusions on the balances: http://msystems.asm.org/content/2/1/e00162-16