Rarefaction and Normalization in Microbiome Analysis
There has been a lot of (sometimes heated) discussion over the years about the pros and cons of various normalization approaches for analyzing microbiome samples sequenced to different depths when using 16S sequencing.
There are three main schools of thought when it comes to how to analyze this data: rarefaction, proportional normalization, and doing nothing.
Doing nothing
The fundamental challenge is that when you sequence a sample more deeply, you increase the likelihood of detecting false positive classifications that are purely the result of sequencing error. In other words, spurious errors in the sequencing reads lead to the read looking more similar to the 16S sequence of a near neighbor to the bug that is actually in you sample.