Molecular descriptors modelling

Let us now extend our molecular descriptor model introduced in Chapter 4 (Eqs. 4-26 and 4-27) to the aqueous activity coefficient. We should point out it is not our principal goal to derive an optimized tool for prediction of yw, but to develop further our understanding of how certain structural features determine a compound s partitioning behavior between aqueous and nonaqueous phases. Therefore, we will try to keep our model as simple as possible. For a more comprehensive treatment of this topic [i.e., of so-called linear solvation energy relationships (LSERs)] we refer to the literature (e.g., Kamlet et al., 1983 Abraham et al., 1990 Abraham, 1993 Abraham et al., 1994a and b Sherman et al., 1996). 

FFBPNN Does not make any assumption of the type of relationship between target property and molecular descriptors Models are difficult to interpret Difficult to design an optimal architecture Risk of overfitting... 

Let us start with a classic example. We had a dataset of 31 steroids. The spatial autocorrelation vector (more about autocorrelation vectors can be found in Chapter 8) stood as the set of molecular descriptors. The task was to model the Corticosteroid Ringing Globulin (CBG) affinity of the steroids. A feed-forward multilayer neural network trained with the back-propagation learning rule was employed as the learning method. The dataset itself was available in electronic form. More details can be found in Ref. [2]. 

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,... 

D descriptors), the 3D structure, or the molecular surface (3D descriptors) of a structure. Which kind of descriptors should or can be used is primarily dependent on the si2e of the data set to be studied and the required accuracy for example, if a QSPR model is intended to be used for hundreds of thousands of compounds, a somehow reduced accuracy will probably be acceptable for the benefit of short processing times. Chapter 8 gives a detailed introduction to the calculation methods for molecular descriptors. 

Molecules are usually represented as 2D formulas or 3D molecular models. WhOe the 3D coordinates of atoms in a molecule are sufficient to describe the spatial arrangement of atoms, they exhibit two major disadvantages as molecular descriptors they depend on the size of a molecule and they do not describe additional properties (e.g., atomic properties). The first feature is most important for computational analysis of data. Even a simple statistical function, e.g., a correlation, requires the information to be represented in equally sized vectors of a fixed dimension. The solution to this problem is a mathematical transformation of the Cartesian coordinates of a molecule into a vector of fixed length. The second point can... 

Kovatdieva A, Golbraikh A, Oloff S, Feng J, Zheng W, Tropsha A. QSAR modeling of datasets with enantioselective compounds using chirality sensitive molecular descriptors. SAR QSAR Environ Re. 2005 16(l-2) 93-102. 

Li H, Yap CW, Ung CY, Xue Y, Cao ZW and Chen YZ Effect of selection of molecular descriptors on the prediction of blood-brain barrier penetrating and nonpenetrating agents by statistical learning methods. J Chem Inf Model 2005 45 1376-1384. 

Our approach is to examine small, closely-related series of nitrosamines and to develop structure-activity models based on molecular descriptors which are explicitly meaningful with respect to the organic chemistry and biochemistry of the compounds. The forms of these models can then often be interpreted in terms of the mechanisms through which these compounds exert their carcinogenic effects. 

An alternative viewpoint for structure-activity investigations is to utilize quantitative models as probes into the mechanism of action of the set of compounds being studied. In this case it is most useful if the molecular descriptors are explicitly meaningful in terms of chemical reactivity or physiological behavior, e.g., distribution of the compound in an organism (see Table II). In a previous symposium, (18), we described our application of this approach toward the development of a quantitative structure-potency expression, equation 1,... 

Numerous other QSAR models relating BBB penetration to calculated molecular descriptors have also appeared in literature see for example [27-29]. In each case, PSA was identified as one of the most important parameters determining blood-brain barrier penetration. 

Hou, T. J., Xu, X. J. ADME evaluation in drug discovery. 3. Modeling blood-brain barrier partitioning using simple molecular descriptors. /. Chem. Inf. Model. 2003, 43, 2137-2152. 

H-bonding is an important, but not the sole, interatomic interaction. Thus, total energy is usually calculated as the sum of steric, electrostatic, H-bonding and other components of interatomic interactions. A similar situation holds with QSAR studies of any property (activity) where H-bond parameters are used in combination with other descriptors. For example, five molecular descriptors are applied in the solvation equation of Kamlet-Taft-Abraham excess of molecular refraction (Rj), which models dispersion force interactions arising from the polarizability of n- and n-electrons the solute polarity/polarizability (ir ) due to solute-solvent interactions between bond dipoles and induced dipoles overall or summation H-bond acidity (2a ) overall or summation H-bond basicity (2(3 ) and McGowan volume (VJ [53] ... 

Calculated molecular descriptors including H-bond parameters were used for QSAR studies on different types of permeabiUty. For example, the new H-bond descriptor characterizing the total H-bond ability of a compound, was successfully appUed to model Caco-2 cell permeability of 17 drugs [30]. A similar study on human jejunal in vivo permeabiUty of 22 structurally diverse compounds is described in Ref. [62]. An exceUent one-parameter correlation of human red ceU basal permeabiUty (BP) was obtained using the H-bond donor strength [63] ... 

More typically the process of building up the QSAR models requires more complex chemical information. For a set of compounds, with known property value, the descriptors are calculated. The process of model building proceeds through a reduction of the molecular descriptors, in order to indentify the most important ones. Then, using these selected chemical descriptors and a suitable algorithm, the model is developed. Finally, the model so obtained has to be validated. 

Given the above-mentioned considerations, it is difficult to believe that it would be possible to fit drugs from all four classes into one single model. However, it is noteworthy that several molecular descriptors do highly influence both permeability and solubility. For example, it has been suggested that the four BCS classes can be divided solely by considering the molecular weight and PSA [16]. 

Tab. 16.5. Molecular hash key descriptor model for the Palm dataset.

The VolSurf method was used to produce molecular descriptors, and PLS discriminant analysis (DA) was applied. The statistical model showed two significant latent variables after cross-validation. The 2D PLS score model offers a discrimination between the permeable and less permeable compounds. When the spectrum color is active (Fig. 17.2), red points refer to high permeability, whereas blue points indicate low permeability. There is a region in the central part of the plot with both red and blue compounds. In this region, and in between the two continuous lines, the permeability prediction is less reliable. The permeability model... 

---