TOOLS

Drug Design Software

For molecular modelling, GeneCode applies an original proprietary methodology that amalgamates all necessary data and computational drivers for the fast and accurate prediction of the properties of compounds and compounds with predetermined properties (Fig.1).

For any target compound, thousands of data points on individual properties and molecular descriptors can be quickly recovered and analysed by using the appropriate internal software drivers.

The molecular descriptor driver can calculate thousands of theoretical characteristics of compounds reflecting their chemical constitution, topological and geometrical structure, electrostatic intra- and intermolecular interactions and a large variety of physical features derived from either ab initio or semiempirical quantum-chemical molecular wave functions.

The model driver encodes a wide diversity of mathematical models for the derivation of the relationship between the experimentally observable properties of compounds including statistical methods such as multi- and nonlinear regression, multivariate methods such as principal component analysis, partial least squares and linear discriminant analysis and machine learning methods such as artificial neural networks, projection pursuit regression, support vector machines and others.

The property driver generates the data fast and dependably for any chemical compound belonging to an applicable class of structures, from the menu of 350 + physical, chemical and biomedically or environmentally important properties (Molcode Toolbox®). The driver also gives a detailed report on the computational model and the descriptors, used together with the statistical evaluation of the reliability of the predicted data.

The compound driver is unique proprietary software for the fast construction of new compounds from large and diverse libraries of molecular fragments, using the original genetic algorithms, pattern recognition techniques and predictive analytics. The new compounds integrate the best (optimum) property values for the given molecular design task. The various data are presented and stored in formats easily transferable between the different drivers and/or external physicochemical and quantum chemical software. It enables quick screening of large external databases (Che MBl Ldb, ZINC, Open NCIDB, PubChem, etc.) with defined compound filters.