Analysis of adaptation through segmented regression with estimation of the join point

Abstract

The methods which relate the individual responses of genotypes to the environment index through simple linear regression equations are not able of identifying the desirable genotype, that is, the genotype which is responsive to favorable or improved environments and maintain reasonable productivity in adverse environments. The segmented linear regression method allows more flexibility for the characterization of the distinct behaviors of the genotype responses to the variation of the environment. However, the arbitrary choice of the null environment index as the change-point of the response rate of the genotypes is a critical aspect of this method. This paper presents a generalization of the segmented linear regression method which considers the change-point as an additional parameter of adaptation. The application of the method is illustrated through an example.