Principal component analysis of 25 metagenomes based on frequencies of COG categories. COG frequencies were normalized to metagenome size. Points are colored by sample type: green = contaminant-degrading microbial consortia, black = waste water/sludge samples, light blue = pristine groundwater and sediment sites, brown = soil samples, yellow = Hawaii Ocean Time Series samples, red = ammonia-oxidizing communities, purple = non-contaminant degrading microbial consortia. See Additional file1: Table S9 for a full list of metagenomes used. All samples are publically available from the JGI IMG-M site (merced.jgi-psf.org/cgi-bin/mer/main.cgi).