Seed morphometry is suitable for apple-germplasm diversity-analyses

The main objective of this study was to evaluate the trustworthiness of seed image analysis as an approach to discriminate apple germplasm accessions. Digital images of seeds from 42 apple cultivars, acquired by a flatbed scanner, provided a phenotypic dataset with 106 morphometric variables. Stepwise Linear Discriminant Analysis (LDA) was used to examine this dataset, and the results were compared with available genetic data. The first comparison among cultivars provided a 38.8% cross-validation of correct identifications with a discriminant percentage ranging between 11.7 and 70%. In agreement with the genetic diversity analysis, the LDA could discriminate between the apples cultivars, identifying two main groups that could be further divided into additional subgroups. Based on our findings, we propose that seed image analysis is a valuable and affordable tool to investigate phenotypic diversity among a large number of apple cultivars.