Chemometric analysis of the multidrug resistance in strains of Penicillium digitatum

Multidrug resistance activities pECr50 of diverse strains of pathogenic fungus Penicillium digitatum against seven toxicants were studied by Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA). Fungal growth data (radii, circumferences, surface areas of fungal colonies, radius differences and ratios) in absence and presence of toxicants were used to derive eight new descriptors for 35 fungal strains. This data set was studied by PCA and HCA, and was correlated with the genome descriptor PCR for expression of gene CYP51 by Partial Least Squares (PLS) regression. Both analyses of pECr50 data and of fungal growth data have identified baseline resistance character, origin and target fruits of the fungal strains. In addition, the analysis of fungal growth data shows that fungal growth morphology is multivariate by nature, meaning that experimental data can be explored more intensely than in usual practice. Fungal growth is directly related to the production of enzyme P45014DM as the main resistance mechanism of P. digitatum against demethylation inhibitors. This is visible from a parsimonious PLS model (two principal components, Q2 = 0.985, R2 = 0.991, SEV = 0.028), validated with eight strains in the external validation set (Q2ev = 0.982, R2ev = 0.990, SECev = 0.025). Chemometric methods in exploring bioassay data are promising approaches to obtain useful information on fungal resistance and to apply these findings in practice.