Differential abundance of genes in surface waters versus those at Station ALOHA at a depth of 4,000 m, based on COG functional categories. All nonredundant proteins predicted in the GOS surface water and the Pacific Ocean 4,000-m-depth WGS data sets were searched against the COG database to assign each protein to a major gene functional category. The differential abundance (y axis) of proteins assignable to each category (x axis) is shown. GOS samples shown are as follows: SAR3 and SAR4 (also identified as GS000_S03 and GS000_S13, respectively), Sargasso Sea (38); GS018, Caribbean open ocean; GS023, Eastern Tropical Pacific open ocean; and GS034, Eastern Tropical Pacific coastal (off the Galapagos islands) (31). COG categories are as follows (adapted from http://www.ncbi.nlm.nih.gov/COG/): A, eukaryotic RNA processing and modification; B, eukaryotic chromatin structure and dynamics; C, energy production and conversion; D, cell division, chromosome partitioning; E, amino acid transport and metabolism; F, nucleotide transport and metabolism; G, carbohydrate transport and metabolism; H, coenzyme transport and metabolism; I, lipid transport and metabolism; J, translation and biogenesis; K, transcription; L, replication, recombination, and repair; M, cell wall/membrane/envelope; N, cell motility; O, protein turnover, chaperones; P, inorganic ion transport and metabolism; Q, secondary metabolism; R, general function prediction only; S, function unknown; T, signal transduction mechanisms; U, intracellular trafficking and secretion; and V, defense mechanisms.

Correspondence between transposase gene content and average Dn/Ds ratio with depth as revealed by fosmid libraries. The average Dn/Ds ratio of a fosmid library (10) relative to the 4,000-m-depth (reference) WGS library (black points, primary x axis) is plotted against the depth that the fosmid library originated from (y axis). Note that the transposase content of the HF libraries (secondary x axis) correlates well to the Dn/Ds ratio of the libraries, particularly from a depth of 10 m to 770 m.

Differential usage of amino acids in protein sequences from surface versus deep waters. Homologs between a surface water WGS library (figure key) and the 4,000-m-depth WGS library were identified and aligned, in a pair-wise fashion, i.e., one sequence from a surface water library against a sequence from the 4000-m-depth library, based on a BLAST approach. The amino acid substitutions at similar (A) and different (B) substitution positions in the pair-wise alignments were compared to provide the differential usage of each amino acid in proteins from a depth of 4000 m relative to ones sampled by each surface water library. Bars represent the n-fold difference (y axes) in the usage of each corresponding amino acid (x axes). The black dotted line represents the fraction of the total amino acid substitution positions evaluated (secondary y axis) for each corresponding amino acid (x axis). For instance, alanine (Ala) constituted the largest fraction of the total substitution positions evaluated among all amino acids (∼10%), whereas tryptophan (Trp) and cysteine (Cys) represented the smallest fraction (constituting ∼1%). Amino acid usage trends were based on a total of about 800,000 substitution positions evaluated for each surface water WGS library. GAP represents the relative abundance of gaps in the alignments of homologous protein sequences.

Codon substitution patterns between proteins from surface waters versus proteins from a depth of 4,000 m. The average Dn/Ds ratio (y axis) for the 4,000-m-depth Pacific Ocean microbial community (filled bars) and the community sampled in the Sargasso Sea sample 4 metagenome (open bars) is plotted against the synonymous substitution rate (Ds) (x axis). Error bars represent one standard deviation of the mean; most means are statistically different by Student's t test. Note that the Dn/Ds ratio is always two to three times higher in the former community, regardless of the time since divergence of the protein sequences under scrutiny (represented by Ds) (29), revealing more relaxed purifying selection at a depth of 4,000 m. Abundant populations of planktonic group I Crenarchaea and Pelagibacter at a depth of 4,000 m also showed two to three times higher Dn/Ds ratios, on average, compared to their surface water counterparts in the Sargasso Sea (inset).

Differential abundance of genes in surface waters versus those at Station ALOHA at a depth of 4,000 m, based on the Pfam database. All nonredundant proteins predicted in the Sargasso Sea sample 4 (SAR4) and the 4,000-m-depth Pacific Ocean WGS data sets were searched against the Pfam database to assign each protein to a Pfam model. The differential abundance (y axis) of proteins assignable to 1,868 Pfam models (x axis), which had at least one significant match in both data sets, is shown. Several important models (discussed in the text) have been annotated. All underlying data of the Pfam analysis, including Pfam models that found significant matches in only one of the data sets, are provided in Table S1 in the supplemental material. Dashed lines represent the log 2 boundary in relative abundance.