Comparison between Bracken outputs

Question

[SushiRoll]

Hey everyone,

I have been doing some taxonomic detection using Kraken2 and later species abundance estimation using Bracken. My input data are short metagenomic reads. My question is the following: If I have 2 different datasets with different read count and I want to compare the abundance in species between these two datasets, should I normalize the Bracken assigned reads to the corresponding total classified reads? I'm confused as to whether Bracken does any sort of normalisation in this sense.

Thanks for any help!