Figure legends
Figure 1 . Illustration of the study design. HTS 10 and HTS 0.5 represent 10 g and 0.5 g (wet weight) treatments for metabarcoding. Sequencing data from metabarcoding were subjected to three different bioinformatics pipelines. Morphological analyses of the diatoms included light microscopy (analyses of 400 valves) and scanning electron microscopy (SEM).
Figure 2. a-c) scatterplots between per-sample richness of taxonomic units from HTS 10 and HTS 0.5 treatments. d) box plot for the number of ESVs/OTUs/morphospecies (log transformed) per treatment. Different letters above the whiskers indicate significant differences according to Kruskal-Wallis pairwise test.
Figure 3. Venn diagrams of the diatom genera relations between HTS 10 and HTS 0.5 treatments (20 corresponding samples).
Figure 4. Venn diagrams of the diatom genera relations between treatments of HTS 10, HTS 0.5 and microscopy (11 corresponding samples).
Figure 5. Mantel correlation plots between HTS 10 and HTS 0.5 treatments.
Figure 6. Mantel correlation plots between microscopy data and metabarcoding treatments (HTS 10 and HTS0.5).
Figure 7. Ordination plots of diatom communities for all treatments. a-d) plots for corresponding 11 samples between microscopy and HTS 0.5 data; e-h) plots for corresponding 14 samples between microscopy and HTS 10 data. Vectors on the plots denote most important variables as based on Random Forest modelling and DistLM marginal tests. Length of the vector represents the importance of variable.