MULTIVARIATE STATISTICAL METHODS AS A TOOL IN THE SELECTION OF ANTAGONIST STRAINS FOR BIOLOGICAL CONTROL OF PLANT DISEASES

The purpose of the study is to demonstrate the capabilities of multivariate statistical methods for redu-cing the labor intensity of selecting antagonist strains for biological protection of agricultural plants from diseases using the example of searching for antagonist strains against fungal diseases of rapeseed (Brassica napus). The antibiotic activity of 9 strains of Bacillus spp was studied using the counterculture method and 1 strain of Streptomyces hygroscopicus against 9 strains of Fusarium spp., 2 strains of Alternaria spp. and 2 strains of Sclerotinia sclerotiorum, which are causative agents of fungal diseases of rapeseed. The width of the growth inhibition zone was used as an indicator. Using discriminant analysis methods, it was established that strains of antagonist bacteria differ statistically significantly (p < 0.001) in the spectrum of antibiotic activity against phytopathogenic fungi, and strains of phytopathogenic fungi, in turn, differ statistically significantly (p < 0.001) in the spectrum of sensitivity to bacterial strains -antagonists. A matrix of correlations was constructed between the sensitivity of different strains of phytopathogenic fungi to a set of antagonist strains. Factor analysis of this correlation matrix showed that the variation in the set of studied strains of phytopathogenic fungi in sensitivity to a set of antagonist strains is 80.3 % explained by the action of two factors with eigenvalues above 1. Based on the factor loadings, it was concluded that factor 1 represents antibiotic substances active against Fusarium spp., and factor 2 represents antibiotic substances active against Alternaria spp. and S. sclerotiorum. This made it possible to reduce the number of test cultures for the search for future antagonists to 2 strains of phytopathogenic fungi, which have maximum factor loadings for factor 1 and factor 2, respectively. This also made it possible to optimize the combination of antagonist strains for the creation of future biological products, combining strains with the maximum value of factor 1 with strains with a maximum factor value of 2.