PMC

Genome-wide expression profiles of Pelagibacter-related populations, in all four depths. X-axis shows the arbitrary numbering of ORFs along the genome of Pelagibacter strain HTCC7211. Y-axis scale represents normalized cDNA to DNA ratio (normalized expression level; see Supplementary Methods) for each ORF. Each colored circle in the stem plot represents a given ORF at a given depth.

Clustering of all cDNA and DNA data sets based on relative abundance of SEED subsystems. Only the most abundant subsystems that together recruited 95% of all reads are displayed. Hierarchical clustering of four DNA and four cDNA samples were performed with euclidean distance and single-linkage method using MATLAB. Color scale represents the proportion of reads assigned to SEED categories relative to the total library size in each sample. Blue to red color indicates low to high representation of SEED categories.

Community-level gene expression profiles based on the GOS protein family database. Cluster-based expression ratio was defined as representation of each GOS cluster in the cDNA library normalized by its representation in the DNA library. GOS clusters that recruited only cDNA reads were arbitrarily set a value of 1 copy of DNA read, to avoid a denominator of 0. (a) GOS clusters were ranked by their cluster-based expression ratios for four depths; (b) The most highly expressed GOS clusters with known or predicted functions were highlighted for each depth.

Three-way comparison of representation of genomes and genome fragments (fully sequenced fosmids) in DNA and cDNA data sets. The 75 m and 125 m data sets were combined as they were the most similar. Each dot represents a genome (fragment), and its proximity to a vertex reflects the enrichment of the corresponding genome (fragment) in the respective sample. Only genomes recruited >0.1% of total reads are displayed. Station ALOHA fosmids represent fosmid sequences that were reported by (). See Supplementary Methods for detail.

Taxonomic classification based on 16S rRNA-bearing reads in DNA and cDNA data sets. Taxonomic assignments were binned at the order level, using the Hugenholtz taxonomy of Greengenes (see Supplementary Methods). 16S rRNA sequences that could not be classified were excluded from the analysis. Y-axis scale represents the percentage of the total classified 16S rRNA reads. Only taxa that represented ⩾1% of all classified reads are displayed. Also note here that, as no replicate data were available for each sample, error bars were absent and thus no statistical inference could be made from the figure.