ANCOM Significance value

Hi, It´s my first time using this Qiime2 pipeline and right now I’m trying to find the significance level used when running ANCOM. I have read that there is no p-value in the test I can report. Does that mean that the test isn´t statistically valid? Is there a value I can use to report the differential abundance of X feature across my sample groups? What values from the ANCOM output do you use for reporting results?

I´m sorry if it is too basic.


Hi @tacvargas,
Welcome to the forum!

There is no p-value because it’s not really possible to calculate one. You can report the W statistic in place which is the count of how many ratio pairs of that feature are significantly different at a fixed level. The significance level in the Qiime2 implementation of ANCOM is already set at the default <0.05 though I believe this value has changed in ANCOM2 in R. You can also report the means/medians or the differences in means/medians of your comparisons in your publication to give an idea of effect size as well.

1 Like

Thanks ! Really helped me !

1 Like

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