Theoretical analyses of testing efficiency in long-term breeding of poplar

The major goal for long-term poplar breeding can be formulated as maximizing annual progress in Group Merit Gain at a given annual budget （GMG/Y＊）. To evaluate different breeding scenarios, a deterministic simulator BREEDING CYCLE ANALYZER covering the most important aspects （gain, cost, time, technique, and gene diversity） of a full breeding cycle was used. The breeding strategies considered was based on pairwise crossing of the selected breeding population and balanced within family selection for the next breeding population. A main scenario and a number of alternative scenarios within these constraints were evaluated using estimates of the best available inputs for poplars. In focus was a comparison between three different testing scenarios for selecting the parents mated to create future breeding generations, thus selecting based on phenotype, clone test or progeny test. For the main scenario, the highest GMG/Y, and the optimal selection age for clone, phenotype and progeny strategies were 0.7480 %, 0.6989% and 0.4675%; 7, 6, and 11 years respectively. Clone test was best except when heritability was high, plant price was high or total budget was low; phenotype strategy was the second except for the case of extremely low narrow-sense heritability, for which the progeny strategy was a little more efficient than phenotype strategy. GMG/Y was markedly affected by narrow-sense heritability, additive variance at mature age, rotation age, plant-dependent cost, total budget and the time needed to produce the test plants, while diversity loss and recombination cost had rather weak effect on GMG/Y. Short rotation age and cheap testing cost favoured all three testing strategies. Comparably short rotation age, low plant-dependent cost and high total budget seem to promote early selection for progeny strategy.