Acta Scientiarum Polonorum: Hortorum Cultus (Jun 2011)
A METHOD OF EARLY SELECTION OF CUCUMBER GENOTYPES INSENSITIVE TO CHILLING BASED ON DATA MINING
Abstract
The paper presents a procedure to identify promising, chilling-insensitive cucumber genotypes from a pool of 55 breeding lines. Selected genotypes may constitute valuable material in further selection of high-yielding cultivars in moderate climate. The approach is based on determining nitrate (NO3–) content, nitrate reductase activity (NRA) and chlorophyll content (Chl) in cotyledons of cucumbers grown at 12oC. These observations were then used to develop a simple algorithm, which facilitates the ordering of genotypes according to their chilling sensitivity by assigning them ranks on the basis of quartile values (from 1 to 4) of determined traits. From the examined collection, 14 least chilling-sensitive genotypes were selected, i.e. their selection for further breeding carries the lowest risk. Low chilling sensitivity of the above mentioned 14 cucumber genotypes was manifested by high Chl levels and high NRA as well as high NO3– contents, i.e. the sum of quartile ranks ranged from 10 to 12, at a maximum of 12. Then cluster analysis was applied to select genotypes, which possess desirable levels of tested traits. Cluster analysis showed that at the division into two and into three subsets all the 14 genotypes considered promising were found within the same cluster, when these genotypes were divided into more subsets, 13 out of the 14 best genotypes were found in one cluster. The presented method may be used to select the least chilling-sensitive cucumber genotypes also from other collections. Knowing the quartile values calculated on the basis of presented results, the rank of new genotypes characterizing their chilling sensitivity may be estimated, provided that experiments are carried out under conditions similar to those used in this study. Otherwise the application of this algorithm has to be preceded by an in-depth explorative analysis and the determination of new quartile values for the analyzed traits.
