Application of ontologies in computational medicinal chemistry studies and relations between chemical structure and biological activity

Abstract

The quantitative study of structure-activity relationship (QSAR) involves the construction of regression models that relate properties (descriptors) of a set containing chemical substances and biological activity data in relation to one or more biological targets in the body. Data sets manipulated by researchers using QSAR techniques are usually characterized by a small number of instances (samples / compounds) and this makes complex predictive modeling more complex. In this context, inhibitors of the biological target ALK-5 will be analyzed and properties (descriptors) will be calculated for this set of inhibitors so that it is possible to develop an ontology that helps the construction and the obtaining of statistical quality QSAR models. This interest in the use of ontologies is based on the fact that the use of a structuring (organization) of the knowledge of a given domain is of great value for the development of more flexible and reliable computational systems for prediction, making the molecular information of the compounds studied for this target are unified and standardized. For the application of this ontology, an approach will be necessary for the selection of chemical data sets related to the target set through the use of already existing ontologies. (AU)