The efficiency and effectiveness of sampling strategies used to develop a core collection for the Polish spring triticale (×Triticosecale Wittm.) germplasm resources

Triticale (×Triticosecale Wittm.) is a hybrid of wheat and rye, which is bred using conventional plant breeding methods. A core collection is defined as a representative sample of the entire plant genetic resources collection that reflects the diversity in the entire collection. A core collection may simplify management and improve the utilisation of the considered germplasm resources. This paper describes and evaluates the efficiency (in sample representativeness sense) of 50 sampling strategies used to establish core collections of Polish spring triticale germplasm resources. Five fractions of core collections (10%, 15%, 20%, 25% and 30% of the entire collection), two clustering methods (Ward’s and UPGMA) and five sample allocation methods based on agro-morphological quantitative traits were compared for their effectiveness using two indices. The first index refers to the average of absolute differences between means across all of the traits in the core and entire collections relative to the means in the entire collection, MD%. The other index is the average of the absolute differences between variances across all of the traits in the core and entire collections relative to the variances in the entire collection, VD%. The results showed that for the studied spring triticale germplasm collection 1) the core collection including at least 20% of the entire collection should be sufficient to provide good representativeness, 2) two of the five sample allocation methods (the proportional and the D2) were characterized by highest level of sample strategies effectiveness, 3) Ward’s method of cluster analysis enabled us to stratify the entire collection in a way that draws more representative core collections than those created using the UPGMA method.