Premise of the study: The production of banana (Musa spp.; Musaceae) plants is affected by various types of somaclonal variations (SV), including dwarfism. However, methods for specific detection of SV are still scarce. To overcome this, a metabolite-based method for detection of dwarf variants was evaluated.
Methods: The gas chromatography-mass spectrometry (GC-MS) metabolite profile of dwarf banana variants was investigated and compared to that of normal-healthy (N) and cucumber mosaic virus (CMV)-infected plants using principal components analysis and partial least squares discriminant analysis (PLS-DA).
Results: Significant differences among the sample groups were observed in 82 metabolites. Rhamnose was exclusively present in dwarf plants but allothreonine and trehalose were present in all but SV samples. Cellobiose was only detected in N plants, while 45 other metabolites, including methyl-glucopyranoside, allopyranose, lactose, phenylalanine, and l-lysine were detected in all but CMV-infected samples. PLS-DA models were able to detect SV, CMV, and N plants with 100% accuracy and specificity.
Discussion: The GC-MS metabolite profile can be used for the rapid, specific detection of SV at early plant production stages. This is the first metabolite-based characterization and detection of somaclonal variation in plants.
Keywords: Musa; Musaceae; cucumber mosaic virus; metabolomics; partial least squares regression; somaclonal variation.