MODELING BIOCHEMICALS WITH LEADING MOLECULAR CONNECTIVITY TERMS

The descriptive power and utility of molecular connectivity terms deri ved by a trial-and-error procedure from a medium-sized set of 8 molecu lar connectivity indices has been tested on different properties of bi ochemical compounds: amino acids, purines and pyrimidines. The trial-a nd-error procedure is sometimes based on the entire set of molecular c onnectivity indices, and sometimes on a subset composed solely of the best molecular connectivity indices. The present paper reports a model ing which encompasses the melting temperatures of natural amino acids, the water solubility of a heterogeneous class of amino acids plus pur ines and pyrimidines, and three different properties of the DNA-RNA ba ses (U, T, A, G and C): singlet excitation energies Delta E-1 and Delt a E-2, and molar absorption coefficient epsilon(260). The modeling is in every case very satisfactory, both from the point of view of predic tive power suid utility of the equations.