394
P. M. Vassiliev et al.
similar features are selected. The selected feature pairs are united into intermediate “integral” features. The obtained “integral” features are compared against each
other and against the QL features that remain on the list, and the resulting analogous
pairs are united again. This procedure is repeated until the possibility of feature
pairing of both primary and intermediate features is completely exhausted.
12.2.9 IT Microcosm Software Complex
All of the theoretical concepts, methods and rules discussed above are implemented
in IT Microcosm software package for Windows [113], which includes 20 applications and a total of over 58,000 lines of source code.
The IT Microcosm software package makes it possible to solve the following
tasks associated with the calculation of prediction regularities:
• Notation of compound activity levels;
• Translation of the structure of a compound into QL representation and the construction of generalized classes of compounds with the desired activity;
• The calculation of first-level consensus decision rules using four prediction
methods;
• The construction of second-level integral consensus decision rules using three
prediction strategies;
• The construction of third-level integral consensus decision rules by generalizing
the prediction results for all strategies;
• The validation of the obtained decision rules;
• Making mixtures of compounds with any composition in the form of a set of
structural files of a standard format;
• Revealing pharmacophorepatterns that are associated with high activity.
The IT Microcosm software package offers researchers the following possibilities
that are associated with the prediction of the presence or extent of pharmacological
and biological activity of various types in organic compounds with any chemical
structure:
• Prediction of the activity of untested compounds;
• Directed search for highly active drug substances;
• Directed design of novel, highly active drug substances;
• Directed search for highly active multi-target compounds that show several
mechanisms of action at the same time;
• Prediction of the activity of organic compound salts taking the mutual effects of
their components into consideration;
• Optimization of qualitative and quantitative composition of organic compound
salts with the aim of achieving the maximum pharmacological effect;
• Predicting the activity of supramolecular complexes that are formed as a result
of a noncovalent intermolecular interaction between several compounds;
• Optimization of qualitative and quantitative composition of supramolecular
complexes with the aim of achieving the maximum pharmacological effect;
P. M. Vassiliev et al.
similar features are selected. The selected feature pairs are united into intermediate “integral” features. The obtained “integral” features are compared against each
other and against the QL features that remain on the list, and the resulting analogous
pairs are united again. This procedure is repeated until the possibility of feature
pairing of both primary and intermediate features is completely exhausted.
12.2.9 IT Microcosm Software Complex
All of the theoretical concepts, methods and rules discussed above are implemented
in IT Microcosm software package for Windows [113], which includes 20 applications and a total of over 58,000 lines of source code.
The IT Microcosm software package makes it possible to solve the following
tasks associated with the calculation of prediction regularities:
• Notation of compound activity levels;
• Translation of the structure of a compound into QL representation and the construction of generalized classes of compounds with the desired activity;
• The calculation of first-level consensus decision rules using four prediction
methods;
• The construction of second-level integral consensus decision rules using three
prediction strategies;
• The construction of third-level integral consensus decision rules by generalizing
the prediction results for all strategies;
• The validation of the obtained decision rules;
• Making mixtures of compounds with any composition in the form of a set of
structural files of a standard format;
• Revealing pharmacophorepatterns that are associated with high activity.
The IT Microcosm software package offers researchers the following possibilities
that are associated with the prediction of the presence or extent of pharmacological
and biological activity of various types in organic compounds with any chemical
structure:
• Prediction of the activity of untested compounds;
• Directed search for highly active drug substances;
• Directed design of novel, highly active drug substances;
• Directed search for highly active multi-target compounds that show several
mechanisms of action at the same time;
• Prediction of the activity of organic compound salts taking the mutual effects of
their components into consideration;
• Optimization of qualitative and quantitative composition of organic compound
salts with the aim of achieving the maximum pharmacological effect;
• Predicting the activity of supramolecular complexes that are formed as a result
of a noncovalent intermolecular interaction between several compounds;
• Optimization of qualitative and quantitative composition of supramolecular
complexes with the aim of achieving the maximum pharmacological effect;