The efficiency of selection in the early generations of a potato breeding programme

[摘要] The efficiency of selection in the early generations of a potatobreeding programme is examined. A number of unselected potato(Solanum tuberosum L) clones were grown in a glasshouse, from truebotanical seed, and thereafter in the field at two locations (a seedsite and a ware site) for three consecutive years. At each stage, anumber of tuber characteristics were visually assessed and yield plusthe yield components were recorded. Four potato breeders visuallyassessed the produce from each clone after harvesting each yearstrials and were in good agreement as to which clones would be selectedin each environment. Selection for visually assessed characters inboth the glasshouse and first clonal year produced a desirableresponse. However, such selection carried a high cost in terms oflosing clones with commercial potential. Selecting clones for yieldin the seedling and first clonal year was only marginally moreeffective than a random reduction in number of genotypes, whileselection in the second clonal year appeared to be somewhat moreeffective as judged by performance in later generations. Comparisonof a random sample of clones with ones from the same crosses which hadbeen selected at the seedling and first clonal year stage was at bestrandom, with some suggestion, however, of a negative effect.The causes behind the inefficiency of selection were found to becomplex. The inefficiency was ascribed, at least in part, to (i) theinaccuracy of assessment on single plant plots; (ii) the "carry-over"effect of the mother tubers and (iii) selection under a short growingseason.Although there was a formally significant interaction betweenprogenies and environments, the rank of the mean of a cross remainedrelatively consistent over different growing conditions. It was foundthat the mean and square root of the variance obtained from breeders'preference in any of the environments provided a good basis forprediction of the number of clones in each cross which would exceed(or equal) a given target value. The square root of the varianceadded increasingly to the accuracy of the prediction as the targetvalue increased but was never a major component in such predictions.When the predictions were used to provide ranking of the crosses, therank correlations showed good agreement between the differentenvironments and between observed and expected ranks. There was noevidence to suggest that univariate cross prediction for any of theother characters under study would not be effective.A new cultivar is unlikely to be successful simply because of highexpression for a single character, but will rather be an all roundimprovement over cultivars already available. Three methods ofmultivariate cross prediction were therefore examined namelymultivariate probabilities, sum of ranks and the frequency ofgenotypes in a sample that transgress set target values. Thecharacters total tuber weigh t, mean tuber weight, number of tubers andregularity of tuber shape were examined. It was found that a sampleas small as 25 clones provided good predictions (as judged by theobserved frequencies in a larger progeny sample examined in variousenvironments). The best estimates were obtained using multivariateprobabilities based on the means, within progeny variances,. and thephenotypic correlations between variates. The ranking of crossesaccording to these multivariate probabilities provided goodindications of the number of clones which survived selection in anactual breeding scheme.Thus it is suggested that an empirical examination of asub-sample of the progeny from a cross could be used to determine thecrosses which would have the highest probability of producingimproved, potato cultivars. Selection of crosses rather thanindividual clones has several advantages which would favour suchtechniques being used in the early generations of a potato breedingprogramme.Cross prediction based on parental performance also provided anindication of the crosses, .and parents, .which would give the highestfrequency of desirable recombinants. These predictions were not asaccurate as those derived from examination of a sub-sample of progenyfrom each cross. But they would allow an earlier, and hence powerful,method of prediction.

[发布机构] University:University of St Andrews;Department:Biology (School of)