Humann3 calculation of RPK values


I accidently found out that in my *_genefamilies.tsv file, the number of UNMAPPED is greater than the input reads number. Do you know what may cause that? How do Humann calculate RPK for UMAPPED



Hmm, that should not be possible. How are you deriving the input reads number?

The abundance value for UNMAPPED is just the number of unmapped reads, so it’s not technically in RPK units (we imagine them piling up on top of a representative unknown 1-kb gene for the purposes of sum-normalization).