High cryptic species diversity is revealed by genome-wide polymorphisms in a wild relative of banana, Musa itinerans, and implications for its conservation in subtropical China

Abstract

Background: Species delimitation is a challenging but essential task in conservation biology. Morphologically similar species are sometimes difficult to recognize even after examination by experienced taxonomists. With the advent of molecular approaches in species delimitation, this hidden diversity has received much recent attention. In addition to DNA barcoding approaches, analytical tools based on the multi-species coalescence model (MSC) have been developed for species delimitation. Musa itinerans is widely distributed in subtropical Asia, and at least six varieties have been documented. However, the number of evolutionarily distinct lineages remains unknown. Results: Using genome resequencing data of five populations (making up four varieties), we examined genome-wide variation and found four varieties that were evolutionary significant units. A Bayesian Phylogenetics and Phylogeography (BP&P) analysis using 123 single copy nuclear genes support three speciation events of M. itinerans varieties with robust posterior speciation probabilities; However, a Bayes factor delimitation of species with genomic data (BFD*) analysis using 1201 unlinked single nucleotide polymorphisms gave decisive support for a five-lineage model. When reconciling divergence time estimates with a speciation time scale, a modified three-lineage model was consistent with that of BP&P, in which the speciation time of two varieties (M. itinerans var. itinerans and M. itinerans var. lechangensis) were dated to 26.2 kya and 10.7 kya, respectively. In contrast, other two varieties (M. itinerans var. chinensis and M. itinerans var. guangdongensis) diverged only 3.8 kya in the Anthropocene; this may be a consequence of genetic drift rather than a speciation event. Conclusion: Our results showed that the M. itinerans species complex harbours high cryptic species diversity. We recommend that M. itinerans var. itinerans and M. itinerans var. lechangensis be elevated to subspecies status, and the extremely rare latter subspecies be given priority for conservation. We also recommend that the very recently diverged M. itinerans var. chinensis and M. itinerans var. guangdongensis should be merged under the subspecies M. itinerans var. chinensis. Finally, we speculate that species delimitation of recently diverged lineages may be more effective using genome-wide bi-allelic SNP markers with BFD* than by using unlinked loci and BP&P.