Descriptors selection

A common objective in descriptor selection is to define a valid descriptor set, which shows an SAR for compounds with affinity for a given biological target. It has 

Recently, an entropy-based approach has been introduced to compare the intrinsic and extrinsic variability of different descriptors, independent of their units and value ranges. The method was originally introduced in communication theory and is based on Shannon entropy, which calculates descriptor-entropy values using histogram representations. Shannon entropy is defined as  

---

Iteration of the steps, descriptor selection, model building, and model validation in combination with an optimi ation algorithm allows one to select a descriptor subset having maximum predictivity. 

The abbreviation QSAR stands for quantitative structure-activity relationships. QSPR means quantitative structure-property relationships. As the properties of an organic compound usually cannot be predicted directly from its molecular structure, an indirect approach Is used to overcome this problem. In the first step numerical descriptors encoding information about the molecular structure are calculated for a set of compounds. Secondly, statistical methods and artificial neural network models are used to predict the property or activity of interest, based on these descriptors or a suitable subset. A typical QSAR/QSPR study comprises the following steps structure entry or start from an existing structure database), descriptor calculation, descriptor selection, model building, model validation. 

To become familiar with genetic algorithms and their application to descriptor selection... 

Figure 9-31. Descriptor selection using a CA. The upper part of the figure shows the original set of descriptors, below which Is the chromosome, Those genes containing the value 1 are selected for the final set of descriptors, which is given at the bottom.

JM Sutter, SL Dixon, PC Jurs. Automated descriptor selection for quantitative structure-activity relationships using generalized simulated annealing. I Chem Inf Comput Sci 35(I) 77-84, 1995. 

Each of these two QSAR model searches led to pools of several thousands of statistically valid linear equations, expressing the estimate of the Cox2 pICso value as linear combinations of molecular descriptors selected by a Genetic Algorithm (GA) [57,... 

Fig. 5.4 Comparative display of a catalyst-HipHop hypothesis and the pharmacophore field descriptors selected by the minimalist ComPharm overlay-based model. ComPharm key features are pinpointed by arrows, while HipHop feature spheres stand for hydrophobes (light blue) and hydrogen bond acceptors (green).

Wegner, J.K. and Zell, A. Prediction of aqueous solubility and partition coefEcient optimized by a genetic algorithm based descriptor selection method. /. Chem. Inf. Comput. Sci. 2003, 43, 1077-1084. 

Re-calculate medians, re-initialize descriptor selection, and re-partition... 

In the NN method, the property F of the target compound is calculated as an average (or weighted average) of that for its NN in the space of descriptors selected for the model. Different metrics (Euclidian distances, Tanimoto similarity coefficients, etc.), can be used to identify the neighbors. Their number k is optimized using a cross-validation procedure for the training set. 

Obtaining a good quality QSAR model depends heavily on many factors in the approach, particularly on the quality of biological data, descriptor selection, and statistical methods (see Chapter 19 for more details). Given the fact that any QSAR approach has strengths and weaknesses, the careful selection of a specific model, or a combination of models, also needs to be emphasized, and is often specific to the particular application in question. 

As an example, using the best 10 descriptors selected by the Genetic Function Approximation approach (Clark and Westhead, 1996 Forrest, 1993) from 153 descriptors, we were able to construct a Decision Tree model consisting of 5 meaningful descriptors ... 

---