Rarefaction of sequences before MetaPhlan analysis

Question

[dpc]

Hi... Should I do rarefaction analysis before MetaPhlAn analysis of WGS metagenome sequences?

Thanks, DC7

[antonioggsousa]

Hi,

According to one of the authors, yes you should rarefy to even sampling depth. Though there are other opinions on the field. Check the discussion here: https://groups.google.com/forum/#!topic/metaphlan-users/e2tets6NHK8

António

[dpc]

Thanks antonioggsousa... I've already checked the discussion. I also have read this article and this one also. they are articles from Cell and plos computational biology journals. And, I am also getting some articles where rarefaction analysis has been done. So, overall I'm confused. However, Do you know. if I can do this step after the Bowtie2 alignment step or after the profiling step (in MetaPhlAn analysis)?

Thanks, DC7

[antonioggsousa]

Yes, there isn't a consensus. This is one of these things that there is a lot of discussion and you probably will never find a consensus solution.

I'm not familiar with MetaPhlAn, but I would say after the profiling step. I worked (and still work) with 16S rRNA gene amplicon data analysis and data transformation or rarefaction or normalization is done after you map and annotate (or profile) your OTUs/ASVs.

The rarefaction here (or normalization or transformation) is important to reduce the difference between the sequencing depth/coverage per sample if you want to perform statistical comparisons between different samples.

I hope this answers your question,

António

[dpc]

Thanks a lot Antonio. Yes. During analysis with MOTHUR, I also have done that after profiling. But, in MetaPhlAn, I am really getting confused when this step to be performed. And I also have to do statistical analysis here.

DC7