Statistical aspects of essential derivation, with illustrations based on lettuce and barley

The concept of essential derivation was introduced by UPOV in 1991 to refine the scope of breeders' rights. The intention of the essential derivation concept was to confer breeders protection against fraudulent practices in which `new¿ varieties are produced from current, protected ones without a genuine breeding effort. A new variety that has passed the standard UPOV tests for Distinctness, Stability and Uniformity, and therefore is eligible for receiving breeders' rights, can still largely be the same as a currently protected variety for a conglomerate of `essential¿, phenotypic traits. Practical implementations of the essential derivation concept entail the definition of a threshold value for genetic conformity between initial and new, putatively derived varieties, beyond which the breeder of the new variety may be asked to prove the genuine nature of his breeding effort. An attractive option for the definition of threshold values for genetic conformity consists in the use of similarities calculated from molecular marker characterizations of varieties. The use of marker based similarities in essential derivation cases raises a number of predominantly statistical questions, such as how to define a reference population of varieties within which potential essential derivation disputes could occur, how many marker loci to use for a required precision, and how to define a threshold value on the basis of the observed distribution of similarity values. This paper describes the first results from special studies undertaken to answer these questions in lettuce and barley