# QSPR Models

The molecular electronic polarizability is one of the most important descriptors used in QSPR models. Paradoxically, although it is an electronic property, it is often easier to calculate the polarizability by an additive method (see Section 7.1) than quantum mechanically. Ah-initio and DFT methods need very large basis sets before they give accurate polarizabilities. Accurate molecular polarizabilities are available from semi-empirical MO calculations very easily using a modified version of a simple variational technique proposed by Rivail and co-workers [41]. The molecular electronic polarizability correlates quite strongly with the molecular volume, although there are many cases where both descriptors are useful in QSPR models. [c.392]

Furthermore, QSPR models for the prediction of free-energy based properties that are based on multilinear regression analysis are often referred to as LFER models, especially, in the wide field of quantitative structure-activity relationships (QSAR). [c.489]

The establishment of QSAR/QSPR models. This process is explained in more detail in Chapter 8. Good QSAR/QSPR models should be interpretable and guide the further development of a new drug. The computer system PASS prediction of activity spectra for substances) allows to predict simultaneously more than 500 biological activities. Among these activities are pharmacological main and side effects, mechanism of action, mutagenicity, carcinogenicity, teratogenicity, and embryotoxicity [19]. [c.605]

The method of building predictive models in QSPR/QSAR can also be applied to the modeling of materials without a unique, clearly defined structure. Instead of the connection table, physicochemical data as well as spectra reflecting the compound s structure can be used as molecular descriptors for model building, [c.402]

It is important to realize that many important processes, such as retention times in a given chromatographic column, are not just a simple aspect of a molecule. These are actually statistical averages of all possible interactions of that molecule and another. These sorts of processes can only be modeled on a molecular level by obtaining many results and then using a statistical distribution of those results. In some cases, group additivities or QSPR methods may be substituted. [c.110]

Polymer modeling is a fast-growing field. It remains primarily the realm of experts because the preferred methods and limitations of existing methods are still changing, thus requiring the researcher to constantly stay abreast of new developments. Group additivity and QSPR methods have been the mainstay of the field due to the difficulty of alternative methods. However, mesoscale and other bulk simulations are becoming more commonplace. Researchers are advised to first consider what properties need to be computed and to then explore the methods and software packages available for those specific properties. [c.315]

As another example, we shall consider the influence of the number of descriptors on the quality of learning. Lucic et. al. [3] performed a study on QSPR models employing connectivity indices as descriptors. The dataset contained 18 isomers of octane. The physical property for modehng was boiling points. The authors were among those who introduced the technique of orthogonahzation of descriptors. [c.207]

The final group of methods used to calculate net atomic charges does not derive them from the electron density, but rather from the electrostatic potential aroimd the molecule. These mclecular-electrostatic- otential (MEP) derived charges are calculated by least-squares fitting of a set of net atomic charges so that they reproduce the calculated MEPs at a grid of points around the molecule as closely as possible. The CHELP [36] and RESP 137] techniques are well known for ab-initio and DFT calculations and MNDO-ESP [38] or VESPA [39] charges can be derived from semi-empirical calculations. Because MEP-derived charges are designed to reproduce the electrostatic properties of molecules as well as possible, they are inherently attractive for describing physical properties. However, in practice the simple Coulson or Mulliken charges have been used more frequently. MEP-derived charges, however, do occur in many QSPR models as the sums of all the MEP-derived charges on atoms of a given element in the molecule. [c.392]

Molecular dipole moments are often used as descriptors in QPSR models. They are calculated reliably by most quantum mechanical techniques, not least because they are part of the parameterization data for semi-empirical MO techniques. Higher multipole moments are especially easily available from semi-empirical calculations using the natural atomic orbital-point charge (NAO-PC) technique [40], but can also be calculated rehably using ab-initio or DFT methods. They have been used for some QSPR models. [c.392]

Juts et al. developed QSPR models for the prediction of solubihty using multiple linear regression analysis (MLRA) and computational neural networks (CNN) (mainly back-propagation neural networks), relating it to the structures of a diverse set of 332 compounds [21]. A series of topological, geometric, and electronic descriptors were calculated. Genetic algorithm and simulated annealing routines, in conjunction with MLRA and CNN, were used to select subsets of descriptors that relate accurately to aqueous solubility. Nine descriptors, including four topological, one geometric, one electronic, and three polar surface area ones, were selected. The model had the corresponding root mean square (RMS) errors of 0.394, 0.358, and 0.343 for the training set, cross-validation set, and test set, respectively. [c.497]

In general, a QSPR/QSAR study starts from a structure database. The molecular structitrc of each compound is entered and stored, providing information about -at least - the molecule s topology (suitable formats are discussed in Sections 2.4 and 2.9. If molecular descriptors are derived from the compound s 3D structure, both experimental and calculated geometries are used. Calculated geometries are submitted to a conformational analysis in order to restrict the study to low-cncrgy conformations. Based on the structure database, a variety of descriptors can be calculated. Optional descriptor subsets are selected. Statistical methods like multilinear regression analysis, or artificial neural networks such as backpropagation neural networks, arc applied to build models. These models relate the descriptors with the property or activity of interest. Finally, the models are validated with an external data set which has not been used for the construction of the model. The steps of a typical QSPR/QSAR study arc summarised as [c.402]

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. [c.432]

Recently, several QSPR solubility prediction models based on a fairly large and diverse data set were generated. Huuskonen developed the models using MLRA and back-propagation neural networks (BPG) on a data set of 1297 diverse compoimds [22]. The compounds were described by 24 atom-type E-state indices and six other topological indices. For the 413 compoimds in the test set, MLRA gave = 0.88 and s = 0.71 and neural network provided [c.497]

See pages that mention the term

**QSPR Models**:

**[c.494] [c.491] [c.54]**

See chapters in:

** Chemoinformatics
-> QSPR Models
**