J. Stejskal

Expansion of the minimum-inbreeding seed orchard design to operational scale

The minimum-inbreeding (MI) seed orchard layout, formulated originally as a global quadratic assignment problem, was expanded into realistic problem sizes that are often encountered in operational forestry, where two modifications were tested: (1) the merging algorithm of independent MI’s solutions (i.e., smaller blocks) and (2) the extended global (genetic-tabu) algorithm. Extending the global heuristic algorithm of the quadratic assignment problem seems to be the most efficient strategy. The reported minimum-inbreeding distance of the extended MI scheme was the lowest in comparison to the completely randomized and the randomized, replicated, staggered clonal-row (R2SCR) seed orchard design schemes. These conclusions also hold for more complex scenarios when added relatedness among orchard’s parents or unequal deployment was considered. This improved MI scheme is suitable to large and complex advanced-generation seed orchards, where many practical constraints have to be jointly considered.

Optimum neighborhood seed orchard design

Original seed orchard algorithm “Optimum Neighborhood Seed Orchard Design” was developed as local heuristics to facilitate the establishment of advanced-generation seed orchards with complex configurations. The scheme leads to uniform spatial distribution among adjacent genetic entries, promoting panmixia. The resulting scheme is thus suitable to establishing both productive seed orchards as well as clonal archives of threatened species, etc. Further, it can accommodate variable clonal sizes, uneven grids, relatedness, assortative mating. It may supplement existing layouts (clonal rows, etc.). Layouts outperformed existing advanced-generation alternative schemes in most parameters and thus could be recommended to operational forestry and gene conservation management.