Characterization of a collection of local varieties of tomato (Solanum lycopersicum L.) using conventional descriptors and the high-throughput phenomics tool Tomato Analyzer

Abstract

Conventional tomato (Solanum lycopersicum L.) descriptors are of great utility for gross morphological characterization but may not be practical for the precise fruit description required for distinguishing closely related cultivar groups. Tomato Analyzer is a new phenomics tool that provides multiple fruit morphology data from scanned images of fruit sections. We characterized 69 accessions of local tomato varieties from the region of València (Spain) corresponding to eight cultivar groups (Borseta, Cherry, Cor, Penjar, Plana, Pruna, Redona, and Valenciana) with 64 conventional and 38 Tomato Analyzer descriptors. Significant differences were found among accessions for all traits except for five monomorphic conventional descriptors, revealing a large diversity in the collection. Significant differences were also found among cultivar groups for 36 conventional and 37 Tomato Analyzer descriptors. The groups Borseta, Cherry, Penjar, Plana, and Pruna were clearly distinct and each of them presented many significant differences with the rest of groups. Conventional descriptors did not differentiate well the Cor, Redona, and Valenciana cultivar groups, but Tomato Analyzer descriptors clearly distinguish Valenciana from Cor and Redona groups. A multivariate principal components analysis (PCA) showed that with the exception of six (8.7 %) accessions, the different cultivar groups (including the very similar Cor and Redona) plotted in separate areas of the PCA graph. The results have shown that combined conventional and Tomato Analyzer descriptors in conjunction with PCA analysis are a powerful tool for characterization and classification of local tomato varieties, as well as for distinguishing between related cultivar groups. This has important implications for the enhancement and protection of local tomato varieties.