# Different Rarefaction Depths Between Sample Types

**URL:** https://forum.qiime2.org/t/different-rarefaction-depths-between-sample-types/24816
**Category:** General Discussion
**Tags:** rarefaction
**Created:** [November 21, 2022, 10:55pm UTC](https://forum.qiime2.org/t/different-rarefaction-depths-between-sample-types/24816 "2022-11-21T22:55:28Z")
**Posts on this page:** 1
**Showing post:** 2

<div class="post-metadata">

### 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: [November 23, 2022, 6:20am UTC](https://forum.qiime2.org/t/different-rarefaction-depths-between-sample-types/24816/2 "2022-11-23T06:20:43Z")

</div>

Hi @cazzlewazzle89 ,

If you have not already, I recommend reading this article (at least figure 1!):

> **[Normalization and microbial differential abundance strategies depend upon...](https://link.springer.com/article/10.1186/s40168-017-0237-y)**
>
> Background Data from 16S ribosomal RNA (rRNA) amplicon sequencing present challenges to ecological and statistical interpretation. In particular, library sizes often vary over several ranges of magnitude, and the data contains many zeros. Although we...

Do _not_ rarefy at different depths, as this will introduce a significant technical variation between sample types. It would be better to not rarefy at all than to rarefy at different depths. If you have significant differences in read counts between samples, you could instead consider using diversity metrics that are insenstive to sampling depths, e.g.:

> [@Robust Aitchison PCA Beta Diversity with DEICODE](https://forum.qiime2.org/t/robust-aitchison-pca-beta-diversity-with-deicode/8333):
>
> DEICODE (pronounced like decode /de.ko.de/) [Documentation available in the pugin library.](https://library.qiime2.org/plugins/deicode/19/) DEICODE is a form of [Aitchison Distance](https://en.wikipedia.org/wiki/Aitchison_geometry) that is robust to high levels of sparsity. DEICODE utilizes a natural solution to the zero problem formulated in recommendation systems called [matrix completion](https://arxiv.org/pdf/0906.2027.pdf). A simple way to interpret the method is, as a robust compositional [PCA (via SVD)](https://en.wikipedia.org/wiki/Principal_component_analysis) where zero values do not influence the resulting ordination. One of the benefits of using DEICODE is the ability to reveal s…

> [@cazzlewazzle89](#):
>
> I agree with the principle of rarefication to avoid biasing diversity by sequencing depth but feel that imposing a very low depth on more diverse samples will undersample them and lose biological information.

This is a very valid concern. Rarefaction would only be needed for classical diversity analyses, however, and not for other steps, e.g., differential abundance testing or qualitative comparisons of taxonomic composition. So the biological information is only lost when estimating alpha and beta diversity. You can use alpha and beta rarefaction methods (i.e., with repeated subsampling, see the actions in q2-diversity) to determine whether the sampling depth is sufficient for making a fair comparison between high- and low-biomass samples, or if the lower rarefaction depth (to enable of low-biomass samples in the comparison) leads to loss of too much information from the high-biomass samples.

The other option of course is to rarefy at different depths if you do not compare the high- vs. low-biomass samples, only make comparisons within these groups (i.e., only compare samples rarefied at the same depth)

Good luck!

---

_[View the full topic](https://forum.qiime2.org/t/different-rarefaction-depths-between-sample-types/24816)._