The understanding of marine microbial ecology and metabolism has been hampered by the paucity of sequenced reference genomes. To this end, we report the sequencing of 137 diverse marine isolates collected from around the world. We analysed these sequences, along with previously published marine prokaryotic genomes, in the context of marine metagenomic data, to gain insights into the ecology of the surface ocean prokaryotic picoplankton (0.1-3.0 mm size range). The results suggest that the sequenced genomes define two microbial groups: one composed of only a few taxa that are nearly always abundant in picoplanktonic communities, and the other consisting of many microbial taxa that are rarely abundant. The genomic content of the second group suggests that these microbes are capable of slow growth and survival in energy-limited environments, and rapid growth in energy-rich environments. By contrast, the abundant and cosmopolitan picoplanktonic prokaryotes for which there is genomic representation have smaller genomes, are probably capable of only slow growth and seem to be relatively unable to sense or rapidly acclimate to energy-rich conditions. Their genomic features also lead us to propose that one method used to avoid predation by viruses and/or bacterivores is by means of slow growth and the maintenance of low biomass.
Molecular taxonomy and phylogeny1 revitalized the field of marine microbiology, allowing for the first time the realization that the 'unseen' and 'unknown' majority of uncultivated microbial taxa could be identified by their 16S ribosomal RNA genes, and identifying widespread clades of marine bacteria and archaea that had no cultivated representatives 2 . More recently, metagenomics has revealed the extent of diversity not fully explained using the available genomes of cultivated marine microbes. When the Sorcerer II Global Ocean Sampling (GOS) expedition metagenomic data were published 3,4 , it was remarkable in its duality of diversity. On the one hand, there seemed to be only a few taxonomic groups that appeared routinely in marine surface water samples. Of these groups, only three (Pelagibacter, Prochlorococcus and Synechococcus) were represented by cultivated marine microbes. On the other hand, despite their apparent ubiquity and abundance in the surface ocean, there was virtually no complete or near-complete genomic assembly of any of the cosmopolitan taxa, implying that these groups, as judged by 16S rRNA sequence, are internally extremely diverse. The GOS data set was also remarkable because of its apparent lack of relatedness to the entire collection of sequenced genomes: using a process called fragment recruitment, which is akin to in silico DNA hybridization, only a few genomes from the entire repertoire of sequenced genomes were found to have a significant number of GOS reads assigned to them.Here we use a large collection of sequenced marine genomes and metagenomic data to show the presence of two major groups in the marine surface picoplankton with striking differences in their met...