Marker-assisted acquisition and core collection formation: A case study in barley using AFLPs and pedigree data

Treuren, R. van; Tchoudinova, I.; Soest, L.J.M. van; Hintum, T.J.L. van


A problem that often occurs in deciding which germplasm should be acquired to expand the diversity of a plant genetic resources collection, and which accessions should be included in a core collection, is the lack of proper data. The usefulness of an AFLP-based protocol to assist in acquisition decisions and in core collection formation was examined by using 52 barley cultivars. For validation purposes, pedigree data of the cultivars were used to calculate the 'effective number of origin lines' (n(o1)), a parameter introduced in earlier research that was defined as the number of alleles per locus, not identical by descent, in a set of lines. Two AFLP primer combinations were able to distinguish all 52 cultivars from each other, and to discriminate between spring and winter crop types. Using the year of origin of the cultivars, the historical development of n(o1) showed a stepwise pattern, indicating the periodical release of genetically similar cultivars, alternated by the incorporation of new material. Comparison of AFLP data between cultivars and both their parents was possible in five cases. These comparisons revealed a high likelihood that the correct parents were involved but a rather skewed contribution of parents to offspring, suggesting that backcrossing had been applied. Treating the 25 cultivars that were released before 1980 and played an important role in barley cultivation as a basic collection, and the 27 more recent cultivars as potential candidates for acquisition, n(o1) values generated by a marker-based approach largely followed those using a random approach. Given this poor performance, a marker-based protocol to assist in acquisition decisions was not considered useful for the analysed material. If the 52 cultivars were considered to be the collection from which a core collection had to be selected, the marker-based selection showed much better results compared to a random selection. About half of the total number of origin lines could be captured with a quarter of the collection, indicating the potential utility of AFLPs in core collection formation.