Similarity matrices based on Bray-Curtis distances, dendrograms (complete linkage clustering) and ordination by non-metric multidimensional scaling (MDS) were then obtained by using the PRIMER 5 software (PRIMER-E, Ltd., UK). One-way analysis of similarity (ANOSIM, Primer-E) was performed on the same distance matrix to test the null hypothesis that there was no difference between eukaryotic communities from replicate samples of each condition. Statistics applied to phylogenetic information From the sequencing results, the beta-diversity was studied from the Unifrac distance (fraction of the total branch length in the
phylogeny that is unique to each environment) of each sample. In order to compare eukaryotic communities from the 9 genetic libraries Unifrac (http://bmf2.colorado.edu/unifrac/index.psp; [47]) metrics were used to perform a principal coordinate analysis (PCA). The P-values matrix that compares Dabrafenib supplier each sample to each other sample was also performed
from UNIFRAC metrics. To investigate the relationships between changes in the eukaryote community structure (number of clones affiliated to each OTUs within main phylogenetic groups) and physic-chemical Olaparib mw and biological parameters, we used direct multivariate canonical correspondence analysis (CCA) [48]. In addition to temperature values, UVB radiation, and nutrient concentrations, we considered the abundances of bacteria, picocyanobacteria, viruses, pigmented eukaryotes and heterotrophic flagellates as Guanylate cyclase 2C explanatory variables. CCA was calculated for the T96 h dataset using the Vegan package within the R software (http://cran.rproject.org/). A minimal set of explanatory variables associated with variation in eukaryote community structure was identified, allowing us to
exclude the most redundant explanatory variables. Forward selection was performed to identify environmental variables that could explain a significant portion of the variation in small eukaryote structure (P < 0.05) at T96 h. Eigen values for site scores, biplot and diversity data were plotted to illustrate the associations between these data [49]. Results Initial conditions Biological and chemical parameters At T0, conditions were considered as homogeneous in all experimental bags. The statistical analysis showed no significant difference between experimental bags in terms of biological parameters (i.e. for bacterial, viral and small eukaryote abundances; mean values are presented in Table 2). Table 2 Initial conditions for chemical and biological parameters Chemical and biological parameters in experimental bags at T0 No nutrient addition + Nutrient PO4 μM 0.07 (±0.01) 0.2 (±0.01) NO3 μM 0.24 (±0.04) 0.32 (±0.05) NH4 μM 0.48 (±0.04) 0.44 (±0.005) NO2 μM 0.04 (±0.004) 0.04 (±0.004) Bacteria 106 cell mL -1* 7.6 (±0.19) 7.8 (±0.37) Virus 108 cell mL-1* 1.5 (±0.3) 1.