# ancom on metaphlan2 data
**URL:** https://forum.qiime2.org/t/ancom-on-metaphlan2-data/11413
**Category:** User Support
**Tags:** metaphlan2, ancom
**Created:** [September 6, 2019, 8:54am UTC](https://forum.qiime2.org/t/ancom-on-metaphlan2-data/11413 "2019-09-06T08:54:50Z")
**Posts on this page:** 3
**Page:** 1
### Author:  [@Stef](https://forum.qiime2.org/u/Stef)
#### Post date: [September 6, 2019, 8:54am UTC](https://forum.qiime2.org/t/ancom-on-metaphlan2-data/11413/1 "2019-09-06T08:54:50Z")
Hi,
I have a metaphlan2 profile in percentage and I want to use ANCOM for statistical analyses. I saw there is a q2-metaphlan2 plugin, so I assume I am not the first person trying to use metaphlan2 profiles in qiime2.
My question is now how to deal with the zeros before using ANCOM. If I run qiime composition add-pseudocount I will replace every '0' to a '1' which is too much if the sum per column is 100. I have around 2 MReads/sample originally but a few only 1 MReads. Could I just multiply all hits times 200,000 to get fake counts? Or what would be the right approach?
Thank you so much for your help!
Cheers,
Stef
---
### Author:  [@Nicholas\_Bokulich](https://forum.qiime2.org/u/Nicholas_Bokulich)
#### Post date: [September 6, 2019, 2:13pm UTC](https://forum.qiime2.org/t/ancom-on-metaphlan2-data/11413/2 "2019-09-06T14:13:24Z")
Hi @Stef,
`add-pseudocount` has a `--p-pseudocount` parameter that allows you to adjust what you would like the pseudocount to be. So you could select something like 0.0000001 (or some other small decimal).
That said, I believe ANCOM is intended for count data, and I do not know whether or not relative abundances like you have will be appropriate for this test. 🤷♂️
---
### Author:  [@system](https://forum.qiime2.org/u/system)
#### Post date: [October 7, 2019, 8:13pm UTC](https://forum.qiime2.org/t/ancom-on-metaphlan2-data/11413/3 "2019-10-07T20:13:27Z")
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.