Developing prokaryotic water quality indicators

Despite the importance of prokaryotes in aquatic ecosystem and their predictable diversity patterns across space time, biomonitoring tools relying on are widely lacking. Using metabarcoding, as well other molecular methods, we were able to identify multiple prokaryotic descriptors illustrate reliability advantages environmental assessment. Multivariate statistical machine learning methods combined with variation coefficient overall prevalence taxonomic groups used detect possible biological indicators among for various anthropogenic pressures, i.e. acidification, eutrophication faecal contamination environments. In addition, text mining approaches provide powerful alternatives sequence based status classification source tracking contaminants. While these individual sequencing indicator seem be powerful, alpha beta indices so far minor precision ecological classification. Reasons often non-linear association between beta-diversity gradients indicated by first modeling attempts. Still, our results suggest that limitations reliably describing reference communities developing general robust systems water quality assessment data can overcome extensive training data.