Predicting germination probability of radish (Raphanus sativus L.), Chinese cabbage (Brassica rapa ssp pekinensis), and cabbage (B-oleracea var. capitata L.) seeds via amino acid leakage parameters

The predictability of the seed germination was investigated on radish (Raphanus sativus L.), Chinese cabbage (Brassica rapa ssp. pekinensis), and cabbage (B. oleracea var. capitata L.) using a logistic regression model with amino acid leakage parameters (color intensity and absorbance of seed leachate treated by the ninhydrin reagent) as predictors. Seeds of eight cultivars of naturally-aged and intact radish, seven cultivars of Chinese cabbage, and one cultivar of cabbage were used. The seeds were individually soaked in distilled water and treated with the ninhydrin reagent. A clearly distinct purple color was observed for the treated seed leachate, with intensities varying with the seed quality. Four separate fractions were identified based on their color intensity, colorless, weak, medium, and strong. Nine fractions from the radish seeds, and eleven fractions from the Chinese cabbage and cabbage seeds were identified based on absorbance value of the color. A germination test was conducted with seeds in each fraction identified. The probability of germination was calculated using a logistic regression model based on the color intensity or the absorbance of the seed leachate. For the prediction of germination based on the color intensity of the seed leachate, the percent correct of predicted germination varied from 72 to 99% and from 82 to 97% in the radish and the Chinese cabbage or cabbage cultivars respectively. Also, the percent correct of predicted germination based on the absorbance of the seed leachate varied from 66 to 100% and from 75 to 97% in the radish and the Chinese cabbage or cabbage cultivars, respectively. In conclusion, seed germination was predictable from both the color intensity and absorbance of the seed leachate. However, different cultivars showed differences between the predicted and actual ratio of germination. The Chinese cabbage or cabbage seeds were more applicable to the prediction model than radish seeds, as the relative error of the germination was slightly lower for the Chinese cabbage or cabbage seeds than the radish seeds.