In the past few decades, it has become apparent that most of the microbial diversity on Earth has never been characterized in laboratory cultures. We show that these unknown microbes, sometimes called “microbial dark matter,” are numerically dominant in all major environments on Earth, with the exception of the human body, where most of the microbes have been cultured. We also estimate that about one-quarter of the population of microbial cells on Earth belong to phyla with no cultured relatives, suggesting that these never-before-studied organisms may be important for ecosystem functions.

word count: 250 16 Text word count: 3,312 17 18Importance 94 words. In the past few decades, it has become apparent that most of the 19 microbial diversity on Earth has never been characterized in laboratory cultures. We 20 show that these unknown microbes, sometimes called "microbial dark matter", are 21 numerically dominant in all major environments on Earth, with the exception of the 22 human body, where most of the microbes have been cultured. We also show that about a 23 quarter of microbial cells on Earth belong to phyla with no cultured relatives, suggesting 24 that, if we can discover their novel functions, they might be important to ecosystem 25 functions. 26 27 28 29 30 31 32 33 34Abstract 35 To unequivocally determine a microbe's physiology, including its metabolism, 36 environmental roles, and growth characteristics, it must be grown in a laboratory culture. 37Unfortunately, many phylogenetically-novel groups have never been cultured, so their 38 physiologies have only been inferred from genomics and environmental characteristics. 39Although the diversity, or number of different taxonomic groups, of uncultured clades 40 has been well-studied, their global abundances, or number of cells in any given 41 environment, have not been assessed. We quantified the degree of similarity of 16S 42 rRNA gene sequences from diverse environments in publicly-available metagenome and 43 metatranscriptome databases, which we show are largely free of the culture-bias present 44 in primer-amplified 16S rRNA gene surveys, to their nearest cultured relatives. Whether 45 normalized to scaffold read depths or not, the highest abundance of metagenomic 16S 46 rRNA gene sequences belong to phylogenetically novel uncultured groups in seawater, 47 freshwater, terrestrial subsurface, soil, hypersaline environments, marine sediment, hot 48 springs, hydrothermal vents, non-human hosts, snow and bioreactors (22-87% uncultured 49 genera to classes and 0-64% uncultured phyla). The exceptions were human and human-50 associated environments which were dominated by cultured genera (45-97%). We 51 estimate that uncultured genera and phyla could comprise 7.3 x 10 29 (81%) and 2.2 x 52 10 29 (25%) microbial cells, respectively. Uncultured phyla were over-represented in 53 metatranscriptomes relative to metagenomes (46-84% of sequences in a given 54 environment), suggesting that they are viable, and possibly more active than cultured 55 clades. Therefore, uncultured microbes, often from deeply phylogenetically divergent 56 groups, dominate non-human environments on Earth, and their undiscovered 57 physiologies may matter for Earth systems. 58 59 Metagenomes Metatranscriptomes Count Count Count Count Count 638 639 640

No abstract