Abstract
Fourier transform infrared spectroscopy (FT-IR) was used to obtain ‘holistic’ metabolic fingerprints from a wide range of plants to differentiate species, population, single plant genotype, and chromosomal constitution differences. Sample preparation simply entailed the maceration of fresh leaves with water, and these samples were then dried and analysed by reflectance FT-IR where spectral acquisition was typically 10 s. All samples gave reproducible, characteristic biological infrared absorption spectra and these were analysed by chemometric methods. FT-IR is not biased to any particular chemical species and thus the whole tissue profiles produced measure the total biochemical makeup of the test sample; that is to say it represents a plant phenotype. We show that by simple cluster analysis these phenotypic measurements can be related to the genotypes of the plants and can reliably differentiate closely related individuals. We believe that this approach provides a valuable new tool for the rapid metabolomic profiling of plants, with applications to plant breeding and the assessment of substantial equivalency for genetically-modified plants.
Original language | English |
---|---|
Pages (from-to) | 489-501 |
Number of pages | 13 |
Journal | Metabolomics |
Volume | 3 |
Issue number | 4 |
DOIs | |
Publication status | Published - 01 Dec 2007 |
Keywords
- artificial neural network
- hierarchical cluster analysis
- discriminant function
- Lolium
- principal components
- Triticum