The pitfalls of r<scp>DNA</scp>‐based <scp>AMF</scp> identification: a comparative analysis of r<scp>DNA</scp> and protein‐coding genes

Intragenomic polymorphism of rDNA in arbuscular mycorrhizal fungi (AMF) has been largely overlooked in ecological and taxonomic studies, and the reliability of nuclear rDNA regions for species identification has not been comprehensively evaluated and compared with protein-coding genes. Analysis of rDNA copies from Rhizophagus irregularis strains revealed average intragenomic distances ranging from 0.08% (small subunit) to 6.9% (internal transcribed spacer 2 (ITS2)), with a maximum of 21.1% in ITS1 within strain DAOM 197198. Intragenomic rDNA polymorphisms are widespread throughout the AM fungal phylogeny, as confirmed by single nucleotide polymorphism density analysis and PacBio sequencing of 148 AM fungal cultures representing 44 species. All commonly targeted rDNA loci in ecological and taxonomic studies are polymorphic, with ITS and large subunit being the most variable, leading to paraphyletic clades and misleading phylogenetic interpretations among closely related species. No unique genetic distance threshold can be applied to identify AMF, because none of the examined protein-coding genes or partial rDNA had a barcode gap. However, indicative distance thresholds of 1% for glomalin, 1.1% for RPB1, and 1.7% for H⁺-ATPase provide guidance for species delimitation. This study characterizes the extent of intragenomic rDNA polymorphism in AMF, underscores the taxonomic challenges posed by highly variable loci, and describes a bioinformatics pipeline for recovering rDNA copies.