I ran maaslin2 on some microbial relative abundance data but I am confused about the results. In the significant results table I only see the following taxa that are significantly associated with my variables of interest (control vs treated and pre vs post randomization).
However, when I open the all results table there are two additional taxa that have p-values <0.05 (highlighted in yellow) that are not in the significant results table. Can someone please explain this?
The q-value for maximum significance is the only editable parameter with the flag max_significance. So if it is under 0.25 (by default) it will make the significant results list even if p value is greater than 0.05
My understanding is the q-value is generated from p-values that are below 0.05 for BH FDR method. So I’m not sure I understand how/why it is included in the significant results list.
I would like to follow up here, as I am having a similar situation with my data. I have several taxa that fall within the specified q-value range, but with p-values above a 0.05 threshold. In cases where the p-value is greater than 0.05 but the q-value is below the established threshold, which value makes the most sense to use to determine statistical significance of the findings?
If you want to assess if the relationship is real, a plot will give you more insight than any p/q value. There should be one among the Maaslin2 outputs, look at that first.
It’s important to adjust for multiple comparisons, but if I was reviewing a paper I would be skeptical of results that didn’t even meet the nominal significance threshold. So you should consider both values.