A structure-information approach to the prediction of biological activities and properties

Abstract

The structure-information approach to quantitative biological modeling and prediction is presented in contrast to the mechanism-based approach. Basic structure information is developed from the chemical graph (connection table). The development, beginning with information explicit in the connection table (element identity and skeletal connections), leads to significant structure information useful for establishing sound models of a wide range of properties of interest in drug design. Skeletal branching patterns and valence state definition lead to relationships for valence-state electronegativity and atom or group molar volumes. Based on these important aspects of molecules, both the electrotopological state (E-State) and molecular-connectivity structure descriptors (chi indices) are developed. A summary of QSAR models indicates the wide range of applicability of these structure descriptors and the predictive quality of QSAR models for protein binding, HIV-1 protease inhibition, blood-brain-barrier partitioning, fish toxicity, carcinogenicity risk, structure space for similarity searching, and data mining. These models are independent of three-dimensional structure information and are directly interpretable in terms of structure information useful to the drug-design process.