Does anyone know why the rarefaction format table (CSV) is not exported with its respective trust values, or from where can I get that data that is generated by the program every time a randomization occurs?
Thanks for your help
That’s not exactly a confidence interval, rather it is a (very tiny) boxplot, so the columns in the CSV (
depth-X_iter-Y) indicate the values that are used to generate the boxplot. So for each depth X, there will be Y values which are used (default is 10, but both X and Y are configurable).
For actual confidence intervals on alpha diversity, you can check out the q2-breakaway plugin, although it expects singletons to be possible, so your upstream processing may not allow that (default usage of DADA2 for example).
Hope that helps!
Thank you very much for your answer, do you want to say then that if I choose to use the Deblur option instead of DADA2, I agree to use the q2-breakaway 1 plugin?
Yep! Deblur allows singletons through if it has enough evidence to believe them, and it seems to be enough for q2-breakaway to choose different models per-sample as appropriate.
To view the confidence intervals you can either
metadata tabulate the resulting
SampleData[AlphaDiversity] or just export the contents. You must not use
feature-table rarefy or the output of
core-metrics-* as that will rarefy your table, which breakaway is designed to avoid entirely.
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.