Why classified reads are contaminated and unclassified are clean reads?

[B-1991-ing]

Dear Karken support team,

Why classified reads are contaminated reads? Is it due to the Kraken2 database a database made up of all possible contaminated bacteria? I used the Kraken2 database to annotate all my original reads, 85 - 99% are unclassified, which means that they are uncontaminated?

Summary of unclassified reads in each of my sample.

Best,

Bing