How to deal with unclassified reads in metagenomics taxonomic classification

Question

[liaoyunshi]

Dear All,

Thanks for reading my questions. I am a freshman in metagenomics study.

I have performed Shotgun Metagenomic Sequencing on human fecal samples, and used read-based classifiers (e.g., MetaPhlAn3) to generate the taxonomic profile for subsequen t analyses (Alpha, Beta diversity).

Thus, I wanted to use microbial species abundace profile for such propose. However, it is almost impossible to obtain a taxonomic profile without any unclassified/unassigned proportions. I wonder how to deal with such parts before diversity analyses.

Let's assume I have such taxonomic profiles (relative abundae from MetaPhlAn3) for two samples. I don't know how I should do before I input this data for species-level diversity analyses.

1. Should I remove the Unassigned proportion from the table?
2. If so, should I re-normalize the relative abundance for Species 1-5 in both samples, respectively?

Please forgive my naive questions and thanks for your help!