Analysis of potential duplicates in barley gene bank collections using resampling of microsatellite data

Redundant duplication among putative Nordic spring barley material held at 12 gene banks worldwide was studied using 35 microsatellite primer pairs covering the entire barley genome. These microsatellite markers revealed an average of 7.1 alleles per locus, and a range of 1 to 17 different alleles per locus. Similarity of accession name was initially used to partition the 174 repatriated accessions into 36 potential duplicate groups, and one group containing 36 apparently unique or unrelated accessions. This partitioning was efficient to produce a distribution of mainly small average genetic distances within potential duplicate groups compared to distances from the group of unique accessions. However, comparisons within potential duplicate groups still contained large genetic distances of the same size as distances between unique accessions indicating classification errors. A bootstrap approach based on re-sampling of both microsatellite markers and alleles within marker loci was used to test for homogeneity within potential duplicate groups. The test was used in each group for sequential elimination of accessions with a significantly large average genetic distance to identify a homogeneous group. Such genetically homogeneous groups of two or more accessions were identified in 22 among the 36 potential duplicate groups studied. Results from the genetic analysis of some potential duplicate groups supported previous conclusions based on passport data through inclusion of the historically most-original accession in the genetically homogeneous group. In other potential duplicate groups the apparently most-original accession according to passport data was not included in the homogeneous set of accessions, indicating that this most-original accession does not have duplicate accessions in the group. During the present study the largest average genetic distance accepted in any homogeneous group was smaller than the smallest distance declared significant in any group, with a threshold average genetic distance of approximately 0.14. The results are discussed with respect to the identification of duplicate accessions within potential duplicate groups, as well as the elimination of genetic off types in such groups. Furthermore, large barley gene bank collections may be screened for potential duplicates with genetic distances below the suggested threshold of 0.14.