A Statistical Evaluation of the Kinship of Protein Molecules

Abstract

An index of correlation suitable for comparison of protein compositions is constructed. The index is defined and its limits stated. Upon application to six seed globulins, as evaluated for ten amino acids, the index is 0.973. For the same ten amino acids in twenty-four proteins of varied type and source the index is 0.625. These results indicate a high degree of kinship for "heterologous" proteins and a higher degree of kinship for "homologous" proteins. The significance of this fact in relation to theoretically possible structures of protein, variability in synthesis, qualitative considerations of rate of molecular evolution, actual protein phylogeny, structural regularity between protein molecules; and related concepts, is discussed.