373
12 Consensus Drug Design Using IT Microcosm
software package is designed to aid in the in silico discovery of chemical compounds with a desired pharmacological activity, which can be defined as “high” in
case of positive effects, and “minimal” in case of adverse effects.
This complex prediction methodology is dramatically different from other approaches to structure-activity analysis in that it simultaneously employs:
• a redundant multidescriptor and multilevel representation of the structure of
chemical compounds as a set of fragment descriptors with different physicochemical meanings and varying extents of complexity;
• several classification methods that differ considerably in their mathematical formalism; and
• several decision-making circuits that are conceptual in the results that they yield.
An additional point to emphasize is that in calculating the prediction dependencies,
no feature space reduction is carried out, and no significant variables are selected;
all of the parameters of the object area description are used in the construction of
the separating functions.
A complex methodology of prediction allows the formation of integral consensus decision rules that are context-independent and work stably in extra-large-dimension correlated spaces. The current version of IT Microcosm 5.1 [113] for the
calculation of QSAR dependencies uses a fourth-order consensus.
The adequacy, validity and high accuracy of IT Microcosm have been demonstrated on multiple occasions when predicting various pharmacological activities
of “conventional” structurally diverse [109, 117, 121, 140] and structurally similar
[80, 127, 131, 132, 130] organic compounds as well as “nonstandard” chemical
systems, complex organic salts [105, 123], supramolecular complexes [128, 129]
and substance mixtures [35, 74, 75, 97, 112, 111, 114, 119, 120], including cases
where the synergy of the components of a mixture was taken into consideration [74,
75, 112, 111, 114].
12.2 Theoretical Basis of IT Microcosm
Information technology, in the general meaning of the word, is a package of technological components (devices or methods, for example) that are used by people to
manage information [20].
Information technology Microcosm for predicting the properties of organic
compounds is a package of original theoretical concepts, mathematical methods,
and rules driving computer algorithms and software that allow a calculated estimate
of the properties of a chemical compound based on its structural formula, with the
help of multilevel consensus classification QSAR dependencies [105].
In IT Microcosm, predictions are based on the task of classifying compounds
into two classes: active compounds and inactive compounds. Compounds are called
active if they show a certain level of a given biological activity; this level is pre-set
by the researcher. Inactive compounds do not meet the requirement of this pre-set
12 Consensus Drug Design Using IT Microcosm
software package is designed to aid in the in silico discovery of chemical compounds with a desired pharmacological activity, which can be defined as “high” in
case of positive effects, and “minimal” in case of adverse effects.
This complex prediction methodology is dramatically different from other approaches to structure-activity analysis in that it simultaneously employs:
• a redundant multidescriptor and multilevel representation of the structure of
chemical compounds as a set of fragment descriptors with different physicochemical meanings and varying extents of complexity;
• several classification methods that differ considerably in their mathematical formalism; and
• several decision-making circuits that are conceptual in the results that they yield.
An additional point to emphasize is that in calculating the prediction dependencies,
no feature space reduction is carried out, and no significant variables are selected;
all of the parameters of the object area description are used in the construction of
the separating functions.
A complex methodology of prediction allows the formation of integral consensus decision rules that are context-independent and work stably in extra-large-dimension correlated spaces. The current version of IT Microcosm 5.1 [113] for the
calculation of QSAR dependencies uses a fourth-order consensus.
The adequacy, validity and high accuracy of IT Microcosm have been demonstrated on multiple occasions when predicting various pharmacological activities
of “conventional” structurally diverse [109, 117, 121, 140] and structurally similar
[80, 127, 131, 132, 130] organic compounds as well as “nonstandard” chemical
systems, complex organic salts [105, 123], supramolecular complexes [128, 129]
and substance mixtures [35, 74, 75, 97, 112, 111, 114, 119, 120], including cases
where the synergy of the components of a mixture was taken into consideration [74,
75, 112, 111, 114].
12.2 Theoretical Basis of IT Microcosm
Information technology, in the general meaning of the word, is a package of technological components (devices or methods, for example) that are used by people to
manage information [20].
Information technology Microcosm for predicting the properties of organic
compounds is a package of original theoretical concepts, mathematical methods,
and rules driving computer algorithms and software that allow a calculated estimate
of the properties of a chemical compound based on its structural formula, with the
help of multilevel consensus classification QSAR dependencies [105].
In IT Microcosm, predictions are based on the task of classifying compounds
into two classes: active compounds and inactive compounds. Compounds are called
active if they show a certain level of a given biological activity; this level is pre-set
by the researcher. Inactive compounds do not meet the requirement of this pre-set