Potential quality indicators are present in all three domains of life
Ecological indices for quality assessment rely on indicator organisms associated to the criteria described in the individual methods. To explore potential indicator organisms across the three domains analysed in the present study, three different types of relationships were investigated. The whole biome and individual domain data were both subjected to same criteria to extract ZOTUs of interest as potential indicators. The results from the whole biome dataset were comparable to the combined results of the individual domains (data not shown). The individual domains were therefore selected for detailed analysis to provide the highest taxonomic resolution possible. A selection of the best potential indicators identified across the three domains of life is shown in Figure 3, the full result is displayed in Figure S5-S7.
The first type of indicators was assumed to be associated to a single ecological status category, and extracted by selecting all ZOTUs that were observed in one third of samples in a single category (with a minimum of n = 2). This yielded 54, 2, and 0 ZOTUs for Bacteria ,Archaea and Eukaryota , respectively. ZOTUs representingRhodobacteraceae , Flavobacterium andFerruginibacter were associated to streams of bad and poor ecological status, while ZOTUs representing Pirellulaceae ,Comamonadaceae and Pseudonocardia were associated to those of good and high ecological status (Figure 3a).
The second type of indicators was assumed to be associated to either low (bad-moderate, n = 12) or high (good-high, n = 38) quality ecosystems, under the definition of the WFD which states that all ecosystems must achieve an ecological status of good or higher. ZOTUs were extracted by selecting those that occurred in at least one third of samples in either group. This yielded 60, 0 and 3 ZOTUs for Bacteria ,Archaea and Eukaryota , respectively. Representatives ofFerribacterium and Rhodobacter were found to be associated to streams of lower ecological status, while ZOTUs representingSolibacillus , Sphingomonadaceae , Continenticola and the insect order Diptera were associated with high quality ecosystems (Figure 3b).
The final type of indicator was assumed to have a strong correlation to ecological status across all categories. ZOTUs that were observed in at least half of the samples in the dataset (n = 50) were extracted and a Spearman’s correlation analysis was performed. A total of 546 ZOTUs met the criteria for Bacteria , and additional filtering for -0.5 < rho > 0.5 reduced this to 34 ZOTUs with strong correlation to ecological status. The Archaea andEukaryota did not yield ZOTUs with a strong correlation, but yielded a total of 6 and 14 ZOTUs, respectively. Strong negative correlation between occurrence and ecological status was observed for ZOTUs representing OM60(NOR5) clade (r = -0.61), Rhodobacteraceae(r = -0.51) and Caenogastropoda (r = -0.47), while a strong positive correlatin was observed for ZOTUs representing Calothrix(r = 0.7), Limnohabitans (r = 0.78), Nitrospira (r = 0.71) and Nitrosomonadaceae (r = 0.65) (Figure 3c).