Research and analysis

Using DNA to understand river diatom communities - summary

Published 10 November 2023

Applies to England

1. Chief Scientist’s Group report summary

This project applied recent developments in DNA data analysis to an existing dataset to see if improvements could be made in the match between ecological classifications obtained using light microscopy and those from DNA, and to explore the potential to extract additional information from the DNA data.

1.1 What did the project set out to achieve?

Benthic diatoms (a group of algae that grow attached to the riverbed) are traditionally identified using microscopes and are used to assess the ecological condition of rivers. The Environment Agency, with the Scottish Environment Protection Agency, developed an alternative DNA-based method which characterises small sections of DNA unique to different diatom species to identify the diatom community on the riverbed. When the methods were compared, about 35% of sites sampled gave a different ecological classification. This “mismatch” would make it difficult to interpret change in ecological status over time, potentially affecting decisions about the causes of decline and need for improvement in rivers. Since the original development of the DNA method, new data analysis approaches have become available. We applied some of these to explore whether we could improve the match between microscopy and DNA. We also examined the relationships between DNA and environmental variables to assess the potential to improve the current approach or to create a new metric for the assessment of the wider (diatom and non-diatom) river phytobenthos.

1.2 What methods were tested?

We updated our reference database, used to compare the DNA data and assign a name (taxonomy) to that piece of DNA, to increase the number of diatom species identified from the DNA samples. We then used four different DNA analytical techniques (bioinformatic pipelines) to extract information on the species present and their relative abundance. Results from the four techniques were compared and used to generate classifications of ecological status which were compared to those generated from microscopy. Additional non-diatom phytobenthic organisms were identified in the DNA data and assessed using different models, to see if these newly detected groups would change our view of the relationship between environmental pressure and ecology, with a focus on the biological response to the plant nutrients phosphorus and nitrogen.

1.3 How well did the DNA techniques perform?

Using an updated reference database did not result in a significant improvement in the match with light microscopy classification using any of the tested bioinformatics pipelines. Results were in line with previous attempts to create a “read-across” between traditional microscopy data and that produced by DNA analysis.  The data produced by light microscopy and DNA analysis are fundamentally different, so making direct comparisons of outputs from these methods is, and will remain, difficult. There was no consistent bias in the results, and neither method can be said to be “right” or “wrong” when they produce different results at a specific site.

One of the tested pipelines (DADA2) was selected to analyse the DNA data to feed into modelling, which looked at the predictive power of both diatom and non-diatom taxa and community responses to nutrients. Both diatom and non-diatom models outperformed the model used in the original DNA method. In addition, results suggest that the combination of DADA2 with new modelling approaches may perform as well as, or better than, the equivalent light microscopy model.  Part of this improvement is likely due to the finer taxonomic resolution offered by DNA analysis.

1.4 What are the next steps?

In future, new assessment tools should be developed based on an understanding of the relationship of DNA outputs to environmental pressures, rather than attempting to fit the new data to existing tools. Future work should explore new metrics and new approaches to predicting ecological status using supervised machine learning algorithms. A multi-taxa microbial assessment of the river benthos should be explored as this is likely to give better assessment of the impact of water quality/stressors than a single taxonomic group.

1.5 Publication details

This summary relates to information from project SC210004, reported in detail in the following output:

  • Report: SC210004/R
  • Title: Using DNA to understand river diatom communities
  • Project manager: Dr Jo-Anne Pitt, Chief Scientist’s Group
  • Research contractor: Dr Daniel Read, UK Centre for Ecology and Hydrology

This project was commissioned by the Environment Agency’s Chief Scientist’s Group, which provides scientific knowledge, tools and techniques to enable us to protect and manage the environment as effectively as possible.

Enquiries: [email protected].

© Environment Agency