A strategy to identify representative maize core collections based on kernel properties

Numerous methods of defining core collection have been developed, but their ability to preserve variability and ease of use must be evaluated primarily for emerging genebanks with low resources. This study described an approach to identify representative maize core collections using kernel properties, the most diverse attributes in maize. Five methods were compared, and four of them successfully created a core collection of 150 accessions (11.73% of the overall collection). Core Hunter 3 was chosen as the best method because it had better diversity properties on the core collection constructed, proven by a low mean difference value, a high variance difference value, a high variable rate value, a coincidence rate value of > 80%, and the highest coverage value. Moreover, the trait properties for this core showed homogeneity to the entire collection in almost all kernel characters. The discovery of an effective strategy for constructing a core set from an extensive collection of maize germplasm while preserving the genetic diversity of the original population will simplify the conservation and promote the use of germplasm.