Figure 4. Temperature as an environmental factor driving the size of biodegradation meta-webs at the eight studied sites.

A significant positive correlation (r² ∼ 0.8; P = 4.5e⁻⁴; t-test) was found between the relative percentage of genes encoding enzymes participating in biodegradation steps (DEGgp) based on the total number of genes (to avoid artifacts due to differences in sample size) (Σ_DNA+16SrRNA predictions; see Supplementary Table S5B) and site temperature. The corresponding data for the BM058 and OV011 sites from the Deepwater Horizon oil spill were also included in the correlation analysis. The total number of unique polluting chemicals (initial substrates or intermediates inferred from both DNA and 16S rRNA data) presumptively accepted as substrates for enzymes in each microbial population is shown in brackets. The Shannon index, as a measure of biodiversity, negatively correlates (r² ∼ 0.69; P = 0.0105; t-test) with site temperature, as shown in the inset graph. Note that a positive correlation (r² ≤ 0.78; P ≤ 3.8e⁻³; t-test) was also found when considering only gene content based on 16S rRNA or DNA data sets (Supplementary Fig. S5).