Exploring the potential of multivariate analysis to study the impact of cotton leaf curl disease on yield traits

Abstract

Cotton production in Pakistan seems adequate, but inclined to several biotic and abiotic constraints, especially Cotton Leaf Curl Virus posing a hazardous impact on cotton yield by affecting yield traits, fiber quality and health status of plants. In this study, eight Bt and four Non-Bt cotton genotypes were sown under augmented design in the experimental area of the Department of Plant Pathology, University of Agriculture, Faisalabad. Presence of virus was confirmed by using the graft inoculation technique in greenhouse while natural inoculum was also relied upon for infection under field conditions. Data encompassing plant height, number of bolls, 50 % open bolls, boll weight, number of sympodial and monopodial branches were collected after disease appearance and subjected to multivariate analyses of variance which revealed that all genotypes responded distinctly to Cotton Leaf Curl Disease. Then these data were subjected to linear discriminant analysis where first two linear discriminants explained 61.05% and 23.54% variability respectively, and highlighted plant height and 50% open bolls as maximum variability explaining cotton yield traits in response to disease. Territorial map was constructed between first two linear discriminants which revealed that disease influenced plant height and 50% open bolls of Bt cotton genotypes potentially than the Non-Bt genotypes. Moreover, data for plant height and 50% open bolls were split and subjected to Hypothesis-error (HE) plot which depicted negative correlation between plant height and 50% open bolls and described the results in conformation with territorial map. Hence, these tools can be potentially used to study the impact of disease on crop yield traits and for selection of lines/varieties either for future breeding program or farmer field.