Figure 1.
Geographic location of the sampling sites in the Mediterranean Sea. PolS – polluted site at Genoa Port; PriS – pristine site at the Montecristo Island.
Table 1.
Background parameters collected for all samples.
Table 2.
Overview of raw sequence read output, processing and annotation of metagenomic profiles.
Figure 2.
Taxonomic annotation by MG-RAST.
Pie charts summarizing the combined taxonomic domain information obtained through annotation against the M5NR database. Bar-chart diagrams on the right indicate extent and quality (e-value distribution) of the annotation from each individual database. Similarities below the 1.0.e-05 e-value threshold, generally less than 20% for each individual database, are coloured blue.
Figure 3.
MG-RAST heatmaps representing the phylogenetic diversity of the four samples at the phylum (left) or class (right) level. Differences between PolS1 and PolS2 are more pronounced than between PriS1/PriS2 sample pairs. Red and green colours indicate low and high abundance, respectively.
Figure 4.
Phylogenetic community diversity.
Diversity represented at the class level as determined using GreenGenes. Values reported represent the percentage fraction present in each sample. Gammabacteria are enriched in the polluted samples PolS1/PolS2 whereas Actinobacteria dominate in the pristine samples PriS1/PriS2.
Figure 5.
OTU distribution based on ssu rRNAs.
OTU distribution as determined by GreenGenes. Only the most frequently encountered OTUs representing about 50% of the total are shown (complete list in Data S4). Clear differences are apparent not only between the two sample classes but also between the two polluted samples PolS1 (blue bars) and PolS2 (red bars). Labels indicate the OTU class: α – Alphaproteobacteria, β - Betaproteobacteria, γ - Gammaproteobacteria, Ac – Actinobacteria, Ba – Bacteroidetes (Flavobacteria), Cy – Cyanobacteria, FL – Flavobacteria, Sy – Synechococcophycideae, Chl - Chloroplast, Clo - Clostridia. (Insert) Principal component analysis of sample diversity using the MG-RAST M5NR protein classification. PriS1 and PriS2 nearly coincide, in contrast to PolS1 and PolS2, which are well separated.
Figure 6.
Functional protein annotation of metagenomic profiles. A
. RAMMCAP COG annotation. Only the 20 COG families with the most significant differences (ShotgunFunctionalizeR) between pairs of polluted and pristine samples are shown (complete list in Data S5). B. Top 30 PFAM families showing the most significant differences (ShotgunFunctionalizeR) between the polluted and pristine samples in the COMET annotation profiles (complete list available in Data S5). Values shown represent the relative abundance of each sample with respect to the sample with the highest abundance (normalized to 1).
Figure 7.
Functional annotation at the pathway and category level.
Examples of whole categories (COG) or pathways (KEGG, SEED) showing significant differences between the two sample groups as determined by ShotgunFunctionalizeR. A complete list is available in Data S6.
Figure 8.
Functional annotation using the SEED subsytem definition.
Comparison of metagenomic profiles at A. the SEED subsystem level 2 or B. at the SEED subsystem level 3. Analysis was performed using ShotgunFunctionalizeR with SEED annotation results from MG-RAST. Only a selection of subsystems showing the greatest differences between the two sample groups is shown. The complete list is available in Data S5.