# number of samples for differential abundance analysis

**URL:** https://forum.qiime2.org/t/number-of-samples-for-differential-abundance-analysis/31517
**Category:** General Discussion
**Created:** [September 26, 2024, 10:32am UTC](https://forum.qiime2.org/t/number-of-samples-for-differential-abundance-analysis/31517 "2024-09-26T10:32:27Z")
**Posts on this page:** 3
**Page:** 1

<div class="post-metadata">

### Author: ![Michele94](https://forum.qiime2.org/user_avatar/forum.qiime2.org/michele94/32/17142_2.png) [@Michele94](https://forum.qiime2.org/u/Michele94)
#### Post date: [September 26, 2024, 10:32am UTC](https://forum.qiime2.org/t/number-of-samples-for-differential-abundance-analysis/31517/1 "2024-09-26T10:32:27Z")

</div>

Dear all,

when i was working on my soil 16 amplicon sequencing data (12 samples taken for 2 treatements), by looking at the Pcoa plot of weighted UniFrac distances, i discovered one sample isolated from all the others and noticed this sample has very high levels of phosphorus compared to the others.  
I then tried DA analysis (ALDEx2 in R studio ggpicrust2) on the q2 picrust2 functional annotated data to confirm my hypothesis but i think i cannot perform it becouse it is just one sample and i think i need at least 3 per group. Can anyone suggest me a way to discover differentally abundant OTUs or functional annotations for 1 sample vs all? or other ways to analyze the same data?

thanks in advance,

Michele

i will report my error here just in case:

computing center with all features  
Error in apply(data[, grp2], 1, sd) : dim(X) must have a positive length

---

<div class="post-metadata">

### Author: ![jwdebelius](https://forum.qiime2.org/user_avatar/forum.qiime2.org/jwdebelius/32/9655_2.png) [@jwdebelius](https://forum.qiime2.org/u/jwdebelius)
#### Post date: [September 26, 2024, 2:57pm UTC](https://forum.qiime2.org/t/number-of-samples-for-differential-abundance-analysis/31517/2 "2024-09-26T14:57:00Z")

</div>

Hi @Michele94,

You could potentially analyze the phospherous concentration as continuous (ANCOM-BC will let you use a continuous predictor) and then you won't have 1 group vs many, you'll have everything in a continous distribution.

However, I still think you're going to have issues with 12 samples along a continous distribution. You may be limited in your interence and may need to focus on descriptive analyses.

Best,  
Justine

---

<div class="post-metadata">

### Author: ![Michele94](https://forum.qiime2.org/user_avatar/forum.qiime2.org/michele94/32/17142_2.png) [@Michele94](https://forum.qiime2.org/u/Michele94)
#### Post date: [September 26, 2024, 3:29pm UTC](https://forum.qiime2.org/t/number-of-samples-for-differential-abundance-analysis/31517/3 "2024-09-26T15:29:15Z")

</div>

Ok, i'll try both then. thank you for the suggestion!