Selecting a core germplasm population using molecular markers: a case study with Eucalyptus pellita grown in common garden trials.

Germplasm resources are crucial for revitalizing of the forest industry and plant breeding. Eucalyptus pellita is one of the most widely planted tropical eucalyptus species used for pulp and timber production. Maintaining large field-based germplasm collections for breeding and conservation is expensive. A smaller "core" population, retaining most of genetic diversity, is an efficient strategy. A comparative analysis using the marker-based tools DARwin, CoreFinder, and Core Hunter identified hypothetical subpopulations from 10 to 15% of the source population size, while still capturing the majority of the broader population diversity. Examination of phenotypic data in the source and hypothetical core populations over 34 months allowed comparison of growth and survival, and the effects of mortality on marker-based diversity parameters. After 34 months, the survival in the source population was 88%, while that in the core subsets ranged from 82 to 90%. Populations with higher heterozygosity showed somewhat superior survival, and the core population constructed using CoreFinder displayed marginally higher observed heterozygosity (Ho) and superior growth traits. Population structure and discriminant principal component (DAPC) analyses revealed that the study population was divided into two main subpopulations: New Guinea (NG) and Australia(AUS). The proportion of samples from NG and AUS selected using the three methods differed significantly, with the ratios from CoreF and CoreH closely resembling the source population. Although survival declined over time, the genetic diversity remained largely constant. Since the growth of the NG material was somewhat better than that of the AUS provenances, we advocate modifying the core selections or conducting selections within regions of provenance to ensure that the best breeding materials are represented in the core subpopulations. We explore the practicality of employing marker-based methods to select a core population in genebanks and/or breeding populations that rely on genetic diversity for trait improvement, which are likely to experience ongoing mortality due to inbreeding depression and competition.