QSAR ANALYSIS OF THE ECDYSTEROIDS’ AFFINITY BASED ON 2.5D-SIMPLEX REPRESENTATION OF THE MOLECULAR STRUCTURE

QSAR-studies of such classes of the organic compounds as steroids, are important because they can expand information about optimal ligands, research of the compounds analogical to existing steroids and understanding factors important for binding with receptors. In this study, we represent «2.5D»-SiRMS approach which is an expansion of the simplex representation of the molecular structure considering the labels of chirality. Steroid set of Kramer was used for test study, and set of ecdysteroids with affinity data to ecdysteroid receptor as the main target of the study. «2.5D»-SiRMS allowed to get satisfying QSAR models for both of those tasks. For Kramer set statistical parameters are: R2=0,84: Q2=0,79, RMSE=0,51. For ecdysteroids set statistical parameters are: R2=0,94: Q2=0,79, RMSE=0,44. Those results are comparable or higher than those of most of 3D-QSAR approaches. Functional and structural interpretations of the QSAR-model for ecdysteroid receptor affinity are also given in this study. There was shown that hydrophobic and electrostatic factors are the most important to the affinity. Analysis of structural fragments’ influence also allowed to distinguish some fragments, which provide activity more than other ones