Cultivars in the species Rosa hybrida L. can be classified into five septet groups according to their karyotypes. The objective of this study was to confirm the measurement of the septets using random amplified polymorphic DNA (RAPD) in various rose cultivars. Three species were comprised of four accessions, and 55 cultivars were subdivided into 37 standard and 18 miniature cultivars. The characteristics of the individual septets in three species were as follows: Rosa multiflora (WD1 and WD2) and R. indica (WD4) belonged to septet A and AA, respectively, and R. rugosa (WD3) belonged to septet C. Moreover, 55 rose cultivars were used in order to assess the contribution of the septets. Nineteen RAPD primers were used for an analysis. As a result of the dendrogram obtained from a cluster analysis, three different rose species were clearly allocated in accordance with their septets. Two accessions of R. multiflora (WD1 and WD2) were grouped together. In this study, the septet concept applied to identify the genetic variations of rose cultivars in more detail. Consequently, RAPD was a reasonable tool for the analysis of genetic backgrounds in rose cultivars.
I knew that someday someone would get around to trying to prove Hurst’s septet theory with statistics applied to genetic sequence data.
These folks took his assignments and classifications verbatim which is a surprise, though. I would have expected an approach that sought segregate groupings independently of Hurst’s system that would thus support the theory; or that no segregation would be found thus disproving the theory.
Basically, I’m not really sure that this paper is proof positive of the septet theory. It has some questionable assumptions and assertions.
This paper is a total mess. The more I look at it, the more surprised I am that it ever got published. Just look at their dendrogram (Figure 1). Their division of that into 3 groups doesn’t reflect the actual branching of the tree for one, and then the group that they are calling “Septet C” includes Rosa rugosa, which is correct, but also modern floribundas like Pinocchio, which has NO rugosa blood. Also, their analysis would indicate that Rosa rugosa is more closely related to Pinocchio than Pinocchio is to a modern miniature rose like Lollipop, which makes no sense, and even their two different accessions of R. multiflora don’t group together. The only conclusions we can draw from this paper is, I think, that the marker system they are using does not generate useful information, and the people who reviewed this paper completely fell down on the job.
Joseph, I’m afraid you are right. The methods used here are ancient in biotech terms. They used a 1993 program to build dendrograms, didn’t discuss how it was validated. But more fundamental is use of only a limited number of probes. The septet concept may have some basis in evolutionary history with the caninae in particular where there is a peculiar meiosis and somehow different groups of 7 dominate and others are silenced. Or in comparison of very wide crosses of divergent species. Certainly we see some funny business when crossing things like Austrian Copper into HT germplasm. But if we have a tetraploid of 28 chromosomes, we need more than 19 probes to have a good probability of hitting all the chromosomes, unless we have probes that are already known to be targeted to particular chromosomes in particular septet backgrounds. There is no evidence than any of these RAPDs were identified as targeting anything in an informative way.
I am familiar with use of microsats in insects to supposedly study population ecology. Scoring presence/absence of particular bands based just on size or band intensity is a very subjective process and independent observers may come to very different readings. so the primary basis of this study has little more going for it than a cytologist reading knob and loops of chromosomes, or a classical botanist classifying traits like prickle distribution and petal shape or color.