An informational diversity framework, illustrated with sexually deceptive orchids in early stages of speciation

Reconstructing evolutionary history for emerging species complexes is notoriously difficult, with newly isolated taxa often morphologically cryptic and the signature of reproductive isolation often restricted to a few genes. Evidence from multiple loci and genomes is highly desirable, but multiple inputs require 'common currency' translation. Here we deploy a Shannon information framework, converting into diversity analogue, which provides a common currency analysis for maternally inherited haploid and bi-parentally inherited diploid nuclear markers, and then extend that analysis to construction of minimum-spanning networks for both genomes. The new approach is illustrated with a quartet of cryptic congeners from the sexually deceptive Australian orchid genus Chiloglottis, still in the early stages of speciation. Divergence is more rapid for haploid plastids than for nuclear markers, consistent with the effective population size differential (N_ep < N_en), but divergence patterns are broadly correlated for the two genomes. There are nevertheless intriguing discrepancies between the emerging plastid and nuclear signals of early phylogenetic radiation of these taxa, and neither pattern is entirely consistent with the available information on the sexual cues used by the orchids to lure the pollinators enforcing reproductive isolation. We describe possible extensions of this methodology to multiple ploidy levels and other types of markers, which should increase the range of application to any taxonomic assemblage in the very early stages of reproductive isolation and speciation.