The establishment of 'essential derivation' among rose varieties, using AFLP

In the International Union for the Protection of New Varieties of Plants Act of 1991, mutation is mentioned as one of the mechanisms to obtain an `essentially derived' variety (EDV). For the implementation of the EDV concept in the case of mutation, it is important that the level of genetic relatedness between an initial variety and derived mutant varieties can clearly be distinguished from the level of relatedness between arbitrary pairs of varieties without a derivation relation. Conditions to be fulfilled for such a distinction include enough genetic differentiation in the germplasm pool of interest, sufficiently low levels of genomic sampling error and technical laboratory error and high reproducibility within and between laboratories. In rose, mutants or `sports' are frequently observed during multiplication, making it a suitable crop for studying the possibilities for introduction of the EDV concept in ornamentals. We studied genetic similarities among 83 rose varieties, including 13 mutant groups. Twelve AFLP primer combinations generated 284 polymorphic markers and 114 monomorphic (fixed) bands. Pair-wise Jaccard similarities between original varieties and derived mutants were close to 1.0 (>0.96), whereas all similarities between original varieties were below 0.80, with 75% of the non-mutant similarities even being below 0.50. Values less than 1.0 for similarity among original varieties and their mutants were to a major extent due to scoring errors. Error rates in automated scoring proved to be lower than those in manually scored and transferred data. Experimental errors, even between laboratories, turned out to be very small. On the basis of a consistent and large difference between similarities, relations between an original variety and its mutants can easily be identified and distinguished from relations between original varieties. These results open the way for implementing the essential derivation concept in rose.