Fig. 2

Red Sea prokaryotic diversity. A Mean sequence relative abundances of the top most abundant phyla (over 1%) in the different layers (SURF, DCM, MS, DSL, MD). Proteobacteria has been further divided into the most abundant classes. B Log of odds ratio of the top most abundant phyla plus Proteobacteria classes. A positive value indicates that a phylum has higher probability of being observed in the mesopelagic (mean of MS, DSL and MD), while a negative value indicates a higher probability of appearing in the epipelagic zone (mean of SURF and DCM). Asterisks represent significant odds ratios (p < 0.05) calculated with Fisher’s exact test. C Boxplot representing the alpha diversity index of Shannon. Annotations under each box show richness as number of species (n) and Pilou’s evenness (J). Letters at the top of the figure indicate the different groups according to ANOVA tests on all 3 indexes (H, n, J). D Principal Coordinates Analysis of Bray–Curtis distances of all the 67 samples considered in the study, color-coded by layer. Scaled (zero-centered) environmental and biological variables that have a significant effect on the multidimensional distribution are shown as arrows. The arrow tip indicates the direction where the effect is maximum and the length indicates the magnitude of the effect (Fluor = CTD chlorophyll a fluorescence, DOC = dissolved organic carbon, Temp = temperature, PAb = prokaryotic abundance cells mL⁻¹, %HNA = percentage of HNA cells, PVol = prokaryotic volume µm³, NO₃²⁻ = nitrate µmol L⁻¹, PO₄³⁻  = phosphate µmol L⁻¹, SiO₃²⁻  = silicate µmol L⁻¹, Sal = salinity)