Compound property prediction

Figure 4.3 Tangent delta curve for tread compound property prediction [6].

Direct property prediction is a standard technique in drug discovery. "Reverse property prediction can be exemplified with chromatography application databases that contain separations, including method details and assigned chemical structures for each chromatogram. Retrieving compounds present in the database that are similar to the query allows the retrieval of suitable separation conditions for use with the query (method selection). 

When structure-property relationships are mentioned in the current literature, it usually implies a quantitative mathematical relationship. Such relationships are most often derived by using curve-fitting software to find the linear combination of molecular properties that best predicts the property for a set of known compounds. This prediction equation can be used for either the interpolation or extrapolation of test set results. Interpolation is usually more accurate than extrapolation. 

The pharmacological properties of phenylethylamines that control selfadministration are complex. The effects of phenylethylamines on a variety of pharmacological measures do not appear to predict the reinforcing effects of these drugs, as measured by the cocaine substitution procedure in primates. Specifically, none of the following behavioral effects of these compounds accurately predict the results of self-administration experiments within the phenylethylamine class (Griffiths et al. 1976 Griffiths et al. 

Physical and Chemical Properties. As reported in Section 3.2, the relevant physical and chemical properties of cyanide compounds are known. Certain physical parameters such as octanol/water partition coefficient and soil partition coefficient that are used generally for covalently bound organic compounds to predict environmental fate and transport are neither available nor useful for most of the ionic cyanide compounds. 

The validation data set constitutes 42 PAHs (Table 11) comprising both unsubstituted and substituted compounds with a wide range of physical and chemical properties. Predictive models developed for PAH compounds in the training data set (Fig. 15) were used to predict values of sorption coefficients. All predicted and observed values were regressed, and recorded significant R2 values as shown in Figs. 16 and 17, while the difference between such values are presented in Table 11. 

QSAR modeling has been traditionally viewed as an evaluative approach, i.e., with the focus on developing retrospective and explanatory models of existing data. Model extrapolation has been considered only in hypothetical sense in terms of potential modifications of known biologically active chemicals that could improve compounds activity. Nevertheless recent studies suggest that current QSAR methodologies may afford robust and validated models capable of accurate prediction of compound properties for molecules not included in the training sets. 

In the future, more correlations should be made, which may eventually allow the design of compounds with predicted luminescent properties. Furthermore, we feel the Ferguson model will be widely accepted and may guide subsequent thinking in the design of efficient luminescent ruthenium(II) complexes. 

Descriptors are widely used for efficient retrieval of similar compounds and also for clustering and property prediction (see [3] for a recent review). The task of the descriptor is to represent a compound such that a biologically (or chemically) relevant similarity can be deduced efficiently from the comparison of two descriptors with a computer. The difficulty in developing a descriptor is, therefore, to find a good trade-off between the coverage of important physico-chemical properties... 

Table 9.2 Excited-state properties predicted by quantum chemical calculations for compound 9d (top CIS including six active orbitals bottom CIS including ten active orbitals)...

Chemical structure and property prediction (including drug-likeness) Molecular similarity and diversity analysis Compound or library design and optimization Database mining... 

Support vector machines In addition to more traditional classification methods like clustering or partitioning, other computational approaches have recently also become popular in chemoinformatics and support vector machines (SVMs) (Warmuth el al. 2003) are discussed here as an example. Typically, SVMs are applied as classifiers for binary property predictions, for example, to distinguish active from inactive compounds. Initially, a set of descriptors is selected and training set molecules are represented as vectors based on their calculated descriptor values. Then linear combinations of training set vectors are calculated to construct a hyperplane in descriptor space that best separates active and inactive compounds, as illustrated in Figure 1.9. 

IV. Is Property Prediction Applicable to Real Substances or Just to Ideal Compounds ... 

This is the main easily accessible method for direct prediction of a property from chemical structure alone. Fragment methods view a molecule as composed of specified parts, which contribute individually to the compound property. 

IV. IS PROPERTY PREDICTION APPLICABLE TO REAL SUBSTANCES OR JUST TO IDEAL COMPOUNDS ... 

---