Extended Data Fig. 1
From: Microbial dark matter could add uncertainties to metagenomic trait estimations

Percent of metagenomic reads not classified to any taxon (a), percent of reads classified as bacteria between forested and non-forested ecosystems (b), and percentage of genome equivalents recovered in the ‘bacterial’ metagenomes that were found in the full metagenomes (c). Classification was done using kraken2 with RefSeq genomes for bacteria, archaea, viruses, fungi, and protists as reference databases. Genomes within the metagenomes were quantified by determining mean coverage of 30 single-copy genes using MicrobeCensus. The R2 value and line of best fit in (a) and (c) are from linear regression. Asterisks in (b) indicate significantly higher percentage of reads classified as bacterial in non-forested environments (p < 0.001).