Using eDNA to assess freshwater bacterial diversity along a forest–non-forest gradient in the afrotropics

Abstract

Healthy ecosystems are critical for maintaining ecosystem services and water security; yet many freshwater ecosystems have been subject to environmental degradation. Impacts are often greatest in water-scarce and developing regions, including across much of Sub-Saharan Africa, where many people lack access to basic drinking water. However, environmental monitoring programmes to track ecosystem health are generally lacking across this region due to limited resources and funding. Recent advances in environmental DNA (eDNA) methods offer an increasingly cost-effective and information-rich solution. Here, we explore the potential of eDNA as a tool for ecological monitoring of freshwater ecosystems in Uganda, East Africa. We sampled eDNA to quantify the bacterial diversity of rivers, streams, and swamps across a gradient of human disturbance in and around Kibale National Park, using off-the-shelf sampling methods that require minimal pre-existing infrastructure. We found distinct bacterial communities between intact and degraded habitats, but the bacterial community in rivers converged when flowing through intact forest. We identified several taxa with differential abundances that might serve as potential bioindicators of degraded ecosystems, and showed that a machine learning tool trained on eDNA can accurately differentiate between intact and degraded habitats. Our proof-of- concept study demonstrates the potential of eDNA as a practical and cost-effective biomonitoring tool for freshwater ecosystems in resource-limited regions, including Sub-Saharan Africa. We also highlight the potential benefits of protected forest in modulating bacterial composition in freshwater ecosystems.