Panacea or Nemesis? Re-assessing the reliability of serum-albumin intron 1 for genotyping Western Palearctic water frogs

Water frogs (genus Pelophylax) are one of the most widespread and diverse, but also most invasive amphibians of the Western Palearctic region. As such, Pelophylax studies face the challenge of identifying similar taxa that hybridize in sympatry. For this purpose, the nuclear marker serum albumin intron 1 (SAI-1) has been used for over a decade in Pelophylax. Initially praised for its diagnosticity, notably to discriminate common species such as the pool frog (P. lessonae), the marsh frog (P. ridibundus) and their hybridogenetic hybrid the edible frog (P. esculentus) without sequencing (by amplicon length polymorphism), SAI-1 was later questioned due to misidentifications and doubtful patterns of genetic divergence. In this study, we incorporate an up-to-date multilocus phylogeographic framework spanning the entire Pelophylax diversification, to re-assess the performance of SAI-1 for lineage identification and discovery. We show that SAI-1 sequences discriminate all Palearctic water frog species and most of their phylogeographic lineages, enabling us to map their distributions and identify the genomes of hybridogenetic hybrids. However, the phylogeny of SAI-1 is aberrant and unrepresentative of the evolutionof the genus. In particular, differentiated P. l. lessonae alleles segregating in the Alpine region mimic a species-level divergence that is not recovered by any other marker. Moreover, the indel polymorphism that supposedly distinguishes P. lessonae from P. ridibundus, as well as the main P. ridibundus lineages from the Balkans (P. r. ridibundus vs kurtmuelleri), are not diagnostic across the entire range of these taxa. Hence, SAI-1 is neither the panacea for nor the nemesis of Pelophylax genotyping. Sequencing SAI-1 shall continue to offer a reliable and informative preliminary approach of single-gene barcoding identification of lineages, but analyses without sequencing, and other applications such as phylogenetic and taxonomic inferences, should be avoided.