The Calculate Descriptors Dialog gives full control of atomic, bond, molecular as well as any user-defined fragment descriptors. Codessa is capable of calculating many descriptors based on three-dimensional structures such as "Surface Area and Volume", "Charged Partial Surface Area", and "Iso-density Surface". |
Codessa supports "Constructed Descriptors", using any currently available descriptor. This provides a simple interface to create "normalized" descriptors based on atom count, or any other count-based descriptor. |
Codessa provides navigation and visualization to explore not only the structures, but also atom and bond descriptors, e.g. "Solvent Accessable Surface Area" by atom. |
Fragments are user-defined subset of atoms within a molecule that are connected by bonds. This allows for decriptors that focus on a common functional group, or exclude atoms that hide a property. Fragment Descriptors can be extremely useful when studying intra-molecular properties. Codessa makes it extremely easy to define fragments in terms of connectivity and environment. Once a fragment string is specified, Codessa will automatically find the fragment(s) in every structure. |
Codessa supports a new and very powerful set of descriptors based on isodensity surfaces. Both AMPAC and Gaussian support isodensity surfaces that are mapped with ESP charge values. Codessa automates the generation of surface data during the descriptor calculation phase. These new descriptors are much more accurate than the corresponding atomic sphere-based models. |
As illustrated by the "Best Multilinear Regression" dialog, Codessa supports extensive options for calculating correlations. Because Codessa organizes all data into user-defined sets, the user has full control of the correlation. |
All data is organized into various tables, and easily accessible to the user. This table shows improved results utilizing fragment-based descriptors as well as the new IsoDensity descriptors. |
Codessa provides detailed plots to help identify "outliers" and explore the quality of the correlations. |
