Percentages in relative abundance of 16s rRNA gene

Question

[marc.acfz]

What is the minimum percentage to be considered in a plot of relative abundance of a 16s or metagomic study? I have seen that generally values above 1% are taken, but is it really good to discard taxa below that value? Do taxa with a low relative abundance have a less relevant ecological function than those with a dominant abundance?

[MSRS]

Thank you for the nice question. Suppose in your study, you may get 200 taxa (OTU). Is it possible to visualize them in a single heatmap, barplot, etc! For different types of study, some people only show the statistically significant texa (reference). Through GraPhlAn, you can visualize most of your texa as a phylogenetic tree.

However, for analysis, we need to consider

1. a cutoff relative abundance
2. statistically significant organisms
3. Organisms of interest

Thank you

Answer 1

[Mensur Dlakic]

Do taxa with a low relative abundance have a less relevant ecological function than those with a dominant abundance?

In practice the answer is yes, but a general answer should be it depends. We don't remove anything from consideration from our metagenomic datasets, regardless of relative abundance.

If you have 10-20 species in a habitat, chances are that 5 or so of them will be abundant in double digits, while a couple of them are likely to be < 1% abundant. It would stand to reason that more abundant species are major determinants of everything that happens in that habitat, as they are major consumers, producers and modifiers of resources.

On the other hand, there are habitats with 200+ different species, in which case just about all of them are going to be < 1% abundant. This is to say that 1% cutoff is very arbitrary, though for practical purposes it may have some value in less populated environments.