Comparison of Different Approaches to Establish a Core Collection of Andigena (Solanum tuberosum Group Andigena) Potatoes

摘要

A core collection of germplasm is a representative set of accessions capturing the complete genetic diversity of the entire collection with minimum repetitiveness. In the present study, in order to develop a core set from an entire collection of 740 accessions of Andigena potato (Solanum tuberosum Group Andigena), data were recorded on 21 morphological descriptors during 2009–2010 autumn crops at Jalandhar, Punjab, India. To select accessions for the core set, the conventional hierarchical cluster analysis and PowerCore, which apply the advanced M strategy of heuristic search, were used. In hierarchical cluster analysis, three sampling strategies within clusters were followed, that is, selections based on random numbers, selections based on maximum principal component analysis (PCA) score and random selections from distinct groups formed in a cluster based on Euclidean distances. Comparison of the range, mean, variance, variance difference, variable rate, coincidence rate and Shannon–Weaver diversity index (H′) of the core sets made by different sampling methods showed that in the use of hierarchical cluster analysis, the core set can be best sampled based on the PCA scores. However, comparison of hierarchical cluster analysis and heuristic search (PowerCore) showed that heuristic search had an edge over the hierarchical cluster analysis as the former covered more diversity, and the coverage was 100% for as many as 19 descriptors in the core set. Besides covering the maximum allelic diversity of the entire collection, the core set of 78 accessions made by the PowerCore also preferentially included accessions possessing desired agronomic characters.