Selection of biochemical characters in the breeding for pest and disease resistance: a method based on analogy analysis of chromatographic separation patterns for emitted plant substances

Chromatographic separation patterns of emitted substances are proposed to be used as fingerprints of the defence chemistry in plants. The goal is to recognize the patterns of resistance agents in the chromatograms and to use these patterns for guidance in resistance breeding. This can be done by comparing a reference set of resistant plants with a reference set of nonresistant plants. One problem is the great number of data. A capillary gas chromatogram from a wounded leaf, e.g., may contain more than a hundred component peaks. Such multidimensional comparisons can only be made by a computer. The literature has been searched for computer programs suited for this purpose. The Statistical Isolinear Multiple Component Analysis method (SIMCA) is found to be best suited. It is theoretically estimated that the processing of chromatographic data by the SIMCA‐method can enable the resistance breeder to classify unknown plants as resistant or nonresistant; to discern nonrelevant peaks in the chromatograms; to grade the importance of the resistance variables; and to predict the optimal levels for the different resistance agents. The results indicate that capillary gas chromatography and high pressure liquid chromatography can be adapted to handle the great number of samples needed in a breeding program. These methods together cover the molecular range of importance for resistance.