# ANCOM more/less abundant taxa **URL:** https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096 **Category:** User Support **Tags:** ancom **Created:** [May 9, 2018, 11:01pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096 "2018-05-09T23:01:38Z") **Posts on this page:** 8 **Page:** 1
### Author: ![1112](https://forum.qiime2.org/letter_avatar_proxy/v4/letter/1/ba8739/32.png) [@1112](https://forum.qiime2.org/u/1112) #### Post date: [May 9, 2018, 11:01pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096/1 "2018-05-09T23:01:38Z")
Hello QIIME2 community, I am exploring how to use QIIME2 2018.4 release, and I struggle with interpreting ANCOM output. To my understanding, the output first table shows the features that reject the null hypothesis, which means that they are differentially abundant. But it does not tell which group has higher/lower abundance. The second table shows the percentile of the features in different groups. Are these numbers the number of reads of those features? Then, does the difference in library sizes between the samples/groups (e.g. 3x - 10x difference) affect the interpretation of the percentile as more/less abundant because they are not normalized? The volcano plot x-axis shows the clr mean\_difference. I searched the forum and [here](https://forum.qiime2.org/t/negative-f-statistic-values-in-ancom-output-after-filtering-low-abundant-taxa/3211/3), it is suggested that it is essentially log fold change and indicates increase if it is positive, and decrease if negative. But in this case (attached picture), the second table shows a higher abundance of the feature in control but the volcano plot gives a negative clr mean difference. I tried switching the order of the sample in the metadata file and it still gives the same negative value. My question is how to tell which group has higher/lower abundance besides from using the percentile table in the ANCOM output? I am sorry if the question is too trivial. Thank you very much ![16](https://forum-qiime2-org.s3.dualstack.us-west-2.amazonaws.com/original/2X/1/179b3b4ae7e09aac4ae34915b54aa228ee4fd771.png) ---
### Author: ![thermokarst](https://forum.qiime2.org/letter_avatar_proxy/v4/letter/t/8e7dd6/32.png) [@thermokarst](https://forum.qiime2.org/u/thermokarst) #### Post date: [May 10, 2018, 3:12pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096/2 "2018-05-10T15:12:47Z")
---
### Author: ![thermokarst](https://forum.qiime2.org/letter_avatar_proxy/v4/letter/t/8e7dd6/32.png) [@thermokarst](https://forum.qiime2.org/u/thermokarst) #### Post date: [May 15, 2018, 1:05pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096/5 "2018-05-15T13:05:32Z")
---
### Author: ![thermokarst](https://forum.qiime2.org/letter_avatar_proxy/v4/letter/t/8e7dd6/32.png) [@thermokarst](https://forum.qiime2.org/u/thermokarst) #### Post date: [May 15, 2018, 1:05pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096/6 "2018-05-15T13:05:34Z")
---
### Author: ![Nicholas\_Bokulich](https://forum.qiime2.org/user_avatar/forum.qiime2.org/nicholas_bokulich/32/19937_2.png) [@Nicholas\_Bokulich](https://forum.qiime2.org/u/Nicholas_Bokulich) #### Post date: [May 15, 2018, 1:40pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096/7 "2018-05-15T13:40:35Z")
Hi @1112, Sorry for the delayed response. > [@1112](#): > > To my understanding, the output first table shows the features that reject the null hypothesis, which means that they are differentially abundant Correct > [@1112](#): > > The second table shows the percentile of the features in different groups. Are these numbers the number of reads of those features? Yes. See [this explanation](https://forum.qiime2.org/t/interpreting-values-from-an-ancom-percentile-abundance-table/1497/14) of the percentile abundance results. > [@1112](#): > > does the difference in library sizes between the samples/groups (e.g. 3x - 10x difference) affect the interpretation of the percentile as more/less abundant because they are not normalized? These frequencies are not normalized in the percentile abundances table; however, they are normalized by ANCOM prior to calculating significance so should not influence the results/interpretation. > [@1112](#): > > the second table shows a higher abundance of the feature in control but the volcano plot gives a negative clr mean difference higher in control, lower in treatment, so negative clr makes sense. As indicated in that forum post, which group is assigned as "control" can be arbitrary (maybe alphabetical order?) so I would not worry about this. > [@1112](#): > > My question is how to tell which group has higher/lower abundance besides from using the percentile table in the ANCOM output? use the percentile abundance table as a guide. In this example, that feature is very clearly more abundant in the control group. I hope that helps! ---
### Author: ![Nicholas\_Bokulich](https://forum.qiime2.org/user_avatar/forum.qiime2.org/nicholas_bokulich/32/19937_2.png) [@Nicholas\_Bokulich](https://forum.qiime2.org/u/Nicholas_Bokulich) #### Post date: [May 15, 2018, 1:40pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096/8 "2018-05-15T13:40:37Z")
---
### Author: ![1112](https://forum.qiime2.org/letter_avatar_proxy/v4/letter/1/ba8739/32.png) [@1112](https://forum.qiime2.org/u/1112) #### Post date: [May 21, 2018, 4:29pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096/9 "2018-05-21T16:29:47Z")
Thank you very much! ---
### Author: ![system](https://forum-qiime2-org.s3.dualstack.us-west-2.amazonaws.com/original/3X/2/1/21af5fe23cb6f4579467c66a9ed94e55274ca7bd.svg) [@system](https://forum.qiime2.org/u/system) #### Post date: [June 21, 2018, 10:39pm UTC](https://forum.qiime2.org/t/ancom-more-less-abundant-taxa/4096/10 "2018-06-21T22:39:29Z")
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.