MycorrhizaFinder: an efficient machine learning tool to quantify endomycorrhizal colonisation of real-world roots

Abstract Background and aims Root colonisation by endomycorrhizal fungi has long been one of the most widely used metrics in mycorrhizal studies. However, due to the significant time required to assess colonisation using traditional microscope techniques, studies of colonisation at large scales are impractical. AI-powered approaches may increase output and facilitate ecosystem assessments. Methods We trained an AI-powered tool (MycorrhizaFinder) on field roots from diverse grasslands and heathlands hosting common Northern European plants with a range of arbuscular (AM) and ericoid mycorrhizal (ErM) fungal structures, and dark septate endophytes (DSE), also common in field-sourced roots. We incorporated a user-customized confidence threshold to encourage the user to engage with inevitable morphological ambiguities, in e.g. ErM and DSE. A Macro F 1 statistic was used to assess the tool’s development. Results We provide a sample workflow from root processing and microscope slide scanning to semi-automated model training and performance evaluation. Without human supervision, our automated baseline Macro F 1 is 66% for arbuscular and at 57% for ericoid mycorrhizal colonisation assessment. Conclusion MycorrhizaFinder is user friendly, requires no programming skills and offers flexibility for advanced agronomists or ecologists who wish to train the tool using their own labelled mycorrhizal root datasets, including images acquired from different instruments or staining protocols. This adaptability allows users to customize the model for specific ecosystems or experimental designs. Leveraged with molecular identification and/or functional assessment of fungi, MycorrhizaFinder could support scalable and repeatable monitoring across ecosystems to assess mycorrhizal status and track land-use changes over time.