Hybridization drives evolution of apomicts in Rubus subgenus Rubus: evidence from microsatellite markers

Abstract

One of the main driving forces of evolution and speciation in the highly apomictic subgenus Rubus in central Europe is sexuality in the series Glandulosi . Palaeovegetation data suggest that initial hybridizations took place over different time periods in the two studied regions, and that the successful origin and spread of apomictic microspecies of the series Radula took place over several millennia. Additionally, the cloning and sequencing show that standard evaluations of microsatellite repeat numbers underestimate genetic variability considering homoplasy in allele size.

Is evolution of apomicts driven by the phylogeography of the sexual ancestor? Insights from European and Caucasian brambles (Rubus, Rosaceae)

“…Such a pattern may also be shaped (rather than produced) by environmental selection, particularly by soil type and climate, as was detected for the West European bramble flora exhibiting strong floristic differentiation on the scale of hundreds of kilometres (Haveman, Bijlsma, de Ronde, & Schaminée, ). These findings agree with our previous studies on other bramble species, which revealed some degree of genetic differentiation within the ecologically rather heterogeneous regions of Colchis and Central Europe (Šarhanová et al., ; Sochor & Trávníček, ; Sochor et al., ).…”

“…Considering multiple shared SSR alleles (E/E; Table ), the latter explanation is more likely. It is, nevertheless, somewhat contradictory to the fact that Caucasian polyploids often exhibit only limited formation of reduced megagametophytes (Sochor & Trávníček, ) and are thus expected to serve mainly as pollen donors in hybridizations (Šarhanová et al., ), not as acceptors like in this particular case. Unidirectional hybridization is nevertheless a ubiquitous phenomenon not only in plants and may be hypothetically caused by many pre‐ or post‐zygotic mechanisms (Turelli & Moyle, ).…”

“…We have previously hypothesized that most polyploid Rubus species/taxa originated in the Holocene in the areas of their current occurrence and that their spread may have been driven by human‐mediated changes in landscape beginning only a few millennia ago. Nevertheless, several polyploid taxa must have been formed before the LGM (or any similarly strong environmental event; Sochor et al., ; Šarhanová et al., ). This is supported by our molecular and ENM data.…”

SSR‐seq: Genotyping of microsatellites using next‐generation sequencing reveals higher level of polymorphism as compared to traditional fragment size scoring

“…The mean amount of size homoplasy (see definition above) was similar between the three studied species (44.7%, 45.9%, and 63.5% for D. fascicularis, M. tridens , and O. obtusangulus , respectively; Table ), which is surprisingly high at the species level considering that the degree of homoplasy increases with increasing time of divergence between populations and taxa (Estoup & Cornuet, ). On the other hand, high degrees of size homoplasy were observed in Rubus subgenus Rubus (based on cloning and sequencing of each SSR locus), detecting SNPs at all studied loci within and among the species (Šarhanová et al, ). Vartia et al () screened 16 individuals of Atlantic cod for homoplasy and detected that 71.7% of the analyzed loci carry sequence variation, which represented 32% of all genotypes.…”

“…While using NGS data from different sequencing platforms for SSR marker development in non‐model plant species is now a common practice (Weising, Wöhrmann, & Huettel, ), NGS is very rarely used for SSR scoring. In order to tackle the homoplasy problem and assess FR variation, some authors combined cloning or single‐strand conformation polymorphism and sequencing (e.g., Germain‐Aubrey et al, ; Lia, Bracco, Gottlieb, Poggio, & Confalonieri, ; Ortí, Pearse, & Avise, ; Šarhanová et al, ), but these methods are costly, time‐consuming, and not easily applicable for polyploids. There are first forays among animals (Bradbury et al, ; De Barba et al, ; Vartia et al, ), but comparisons between traditionally scored fragment length data and the information obtained from sequencing are still missing.…”