Dragon is the software for the calculation of molecular descriptors. These descriptors can be used to evaluate molecular structure-activity relationships (QSAR/QSPR), similarity analysis and screening of molecule databases.
| Dragon 6
Input files. Together with the molecule formats allowed in the previous versions of Dragon (MDL, Sybyl, HyperChem, Macromodel and Smiles), new formats are now supported: CML (Chemical Markup Language) and HyperChem format as a unique file. There are no restrictions on the number of molecules and on the atom types. The novel Dragon graphical interface for molecule import is very flexible allowing also selection of structures that are stored by different file formats and located in different folders. Molecular structures stored in files of different types can be loaded and contemporarily processed in the same batch.
Graphics and statistics. Graphical tools are available in Dragon 6 for a preliminary descriptor analysis. These graphs allow a preliminary analysis of molecule distribution in the descriptor space, as well as a preliminary correlation analysis. Look at the screenshots of Dragon 6.
A molecule viewer is provided to display the molecular structures that have been imported for descriptor calculation.
Loading of user-defined variables. Dragon allows the user to import up to 200 variables and string fields that can be added to the calculated molecular descriptors such as experimental properties and information on the processed molecules (e.g., CAS number, product code).
Extended utilities for command line interface. The Dragon command line interface allows the execution of Dragon as a background application. Dragon 6.0 has now all the same facilities when running as command line interface as the graphical interface. For instance, it is now possible to apply the pair correlation criterion to exclude redundant descriptors from the output file. Moreover, the novel script wizard that guides the user in creating the script file can be directly launched by the graphical interface.
Descriptor calculation. In Dragon 6.0, the user is allowed to customize descriptor calculation and choose the atom weighting schemes before descriptor calculation, whether apply logarithmic transformation to spectral moments and walk and path counts. The novel window for descriptor selection allows also selection of single molecular descriptors included in the different blocks.
Extended capability to process a variety of molecular structures. Restrictions on atom types of molecules are now avoided. Therefore, Dragon is now able to deal with any type of connected structure that also includes undefined chemical elements.
Projects and templates. Results of any calculation can be saved into a project that can be stored and later reloaded to continue result analysis and export of different selections of molecular descriptors. Moreover, to make recursive calculation faster, a template can be created, where information on descriptor selection and general options can be saved.
Principal Component Analysis can be calculated on selected sets of descriptors. PCA results can be easily exported or analysed with the graphical tools provided in Dragon.
DProperties: dProperties is an easy-to-use application for the calculation of physicochemical properties and drug-like indices. These properties can be used to characterise molecules, for screening and filtering of molecule databases.
dProperties calculates molecular properties, such as constitutional features (molecular weight, number of atoms, Hydrogen atoms, Carbon atoms, Nitrogen atoms, Oxygen atoms, halogen atoms, sp3 hybridized Carbon atoms), unsaturation count and index, hydrophilic factor, molar refractivity, topological polar surface area, Moriguchi octanol-water partition coefficient, Ghose-Crippen-Viswanadhan octanol-water partition coefficient, Surface areas, McGowan volume, van der Waals volume, packing density index, Verhaar base-line toxicities, and drug-like indices (such as Lipinski Alert Index, Drug-like scores and Lead-like Scores, Unsat index, Dragon consensus drug-like score, property ranges).
Principle Component Analysis: