AUTHOR=George Paul B. L. , Creer Simon , Griffiths Robert I. , Emmett Bridget A. , Robinson David A. , Jones Davey L. TITLE=Primer and Database Choice Affect Fungal Functional but Not Biological Diversity Findings in a National Soil Survey JOURNAL=Frontiers in Environmental Science VOLUME=7 YEAR=2019 URL=https://www.frontiersin.org/journals/environmental-science/articles/10.3389/fenvs.2019.00173 DOI=10.3389/fenvs.2019.00173 ISSN=2296-665X ABSTRACT=

The internal transcribed spacer (ITS) region is the accepted DNA barcode of fungi. Its use has led to a step-change in the assessment and characterisation of fungal communities from environmental samples by precluding the need to isolate, culture, and identify individuals. However, certain functionally important groups, such as the arbuscular mycorrhizas (Glomeromycetes), are better characterised by alternative markers such as the 18S rRNA region. Previous use of an ITS primer set in a nationwide metabarcoding soil biodiversity survey revealed that fungal richness declined along a gradient of productivity and management intensity. Here, we wanted to discern whether this trend was also present in data generated from universal 18S primers. Furthermore, we wanted to extend this comparison to include measures of functional diversity and establish trends with soil types and soil organic matter (SOM) content. Over the 413 individual sites examined (arable, grassland, woodland, moorland, heathland), we found congruent trends of total fungal richness and β-diversity across land uses, SOM class, and soil type with both ITS and 18S primer sets. A total of 24 fungal classes were shared between datasets, in addition to 15 unique to ITS1 and 12 unique to 18S. However, using FUNGUILD, divergent trends of functional group richness became apparent, especially for symbiotrophic fungi, likely driven by an increased detection rate of Glomeromycetes in the 18S dataset. The disparate trends were also apparent when richness and β-diversity were compared to soil properties. Additionally, we found SOM class to be a more meaningful variable than soil type biodiversity for predicting biodiversity analyses because organic matter was calculated for each sample whereas soil type was assigned from a national soil map. We advocate that a combination of fungal primers should be used in large-scale soil biodiversity surveys to capture important groups that can be underrepresented by universal barcodes. Utilising such an approach can prevent the oversight of ubiquitous but poorly described species as well as critically important functional groups.