I Have analyzed metagenomic NGS data and now I have too many bacteria species and most of them has very low percentage like 0.01 and less values. Since I try to write an article and do some pis charts I can't represent all of them which would be so chaotic. So what ahould be the lowest percentage that I need to include in my article ? Is there a threshold for this that is commonly used? Thanks for your answers in advance.
Do you have to present the data at species level? Perhaps you could consider some higher level (genus, family..)? Alternatively, how many species are there if you apply 1% abundance cut off? How many are there at 5% abundance cut off? Presenting more than 50 data points with names in any kind of picture is basically just a big mess. Also, I don't know what you have done, but it's almost certain that many of the rare taxa are just false assignments..
This really depends on what are actually expecting and what was the objective of this study. If you have used MG-RAST for metagenomic study, you will also get eukaryotic taxonomic assignment that might be from the host from which the sample was derived or result of some contamination.
If you are interested in all the taxonomic assignments, there is no threshold as you will miss out important assignment though very low in number. In general, a really low number like say 0.1% would not be significant, hence, even MG-RAST plots the abundance of top 50 taxonomic entities at each level like Phylum, Class, Order, Family and so on.
In your case, there would be samples in which ~90% data would be unassigned, hence, even 0.01% may be significant.
You can choose to take only top 40-50 taxonomic assignments. That may include a very high or very low percentage.
