Assessing variability in eDNA metabarcoding

: implications for ecological interpretation in environmental monitoring

Resumo

Environmental DNA is an invaluable source of ecological and environmental data. It is a field of research which has seen rapid expansion and subsequent adoption into some statutory monitoring schemes. Standardisation of the protocols involved and understanding of variability within eDNA use has been slower. Scottish aquaculture has adopted eDNA metabarcoding into its environmental
impact monitoring, providing a real-world example of eDNA use with standardisation gaps. This thesis focused on eDNA metabarcoding, primarily of the V3-4 region of the bacterial 16S gene. This thesis contributes to standardising eDNA use and knowledge of variability both for Scottish aquaculture and for global eDNA use. It achieves this by: (i) exploring sediment sample stability
between collection and preservation at different temperatures; (ii) exploring sediment sample stability during preservation by freezing; (iii) investigating variability within the BLAST taxonomic assignment process of metabarcoding data intended for environmental impact monitoring; and (iv) comparing eDNA with traditional macrobenthic-based indices during the recovery of a salmon farm site after cessation of farming. The sample stability studies identified sensitivities to both temperature and freeze-thaw cycles. They highlighted key taxa impacted by both, and the resulting effects on ecological interpretation. Together they identify priorities in sample storage and clarify sample viability given challenging circumstances. This can prevent over-conservative sample discard, and it identifies the issues that will occur with incorrect sample storage. The results of the BLAST searches identified the parameters most influential on V3-4 16S metabarcoding results and identified some commonly reported values as insignificant. The BLAST study evidenced that some parameters can create biases depending on the sample source, indicating that they should be used cautiously. The study also showed how some noise reduction methods interact with different BLAST parameters and analysis methods. The recovery of the salmon farm site showed that macrobenthos and bacteria recovered at similar rates, but that the 16S signal was only observable using taxa-sensitive analysis
methods. Macrobenthic recovery was also clearer when using a taxa-sensitive approach. This study also explored 18S and CO1 markers and found they could not provide any clear signal of either impact level or recovery. The robustness and noise tolerance of a trained random forest have been evidenced throughout this thesis. This thesis explored several areas of variability within eDNA
metabarcoding, highlighting potential issues, and strengths and weaknesses of eDNA. This thesis contributes specifically to Scottish aquaculture, as well as informing wider eDNA use both locally and globally.

Data do prémio	16 mar. 2026
Idioma original	English
Instituição de premiação	University of the Highlands and Islands
Promotores	Super DTP
Supervisor	Tom Wilding (Supervisor) & Barbara Morrissey (Supervisor)