Compound structural descriptors

Selecting an optimum group of descriptors is both an important and time-consuming phase in developing a predictive QSAR model. Frohlich, Wegner, and Zell introduced the incremental regularized risk minimization procedure for SVM classification and regression models, and they compared it with recursive feature elimination and with the mutual information procedure. Their first experiment considered 164 compounds that had been tested for their human intestinal absorption, whereas the second experiment modeled the aqueous solubility prediction for 1297 compounds. Structural descriptors were computed by those authors with JOELib and MOE, and full cross-validation was performed to compare the descriptor selection methods. The incremental... 

Inductive methods for establishing a correlation between chemical compounds and their properties are the theme of Chapter 9. In many cases, the structure of chemical compounds has to be pre-processed in order to make it amenable to inductive learning methods. This is usually achieved by means of structure descriptors, methods for the calculation of which are outlined in Chapter 8. 

The information content of a structure descriptor depends on two major factors a) the molecular representation of the compound b) the algorithm which is used for the calculation of the descriptor. 

Structure descriptors can be distinguished by the data type Table 8-1) of the descriptor and the molecular representation of the compound (Table 8-2). 

Another, probably more broadly applicable, technique is to represent a chemical compound by some of its properties. Figure 8-15 is an extension of Figure 8-1 and shows that when no structure descriptors can be derived because the structure is not known, thcji a compound can be represented by a second property (2) or. better, a series of properties, in order to predict the property 1 of interest. 

Molecules can be represented by structure descriptors in a hierarchical manner with respect to a) the descriptor data type, and b) the molecular representation of the compound. 

Aqueous solubility is selected to demonstrate the E-state application in QSPR studies. Huuskonen et al. modeled the aqueous solubihty of 734 diverse organic compounds with multiple linear regression (MLR) and artificial neural network (ANN) approaches [27]. The set of structural descriptors comprised 31 E-state atomic indices, and three indicator variables for pyridine, ahphatic hydrocarbons and aromatic hydrocarbons, respectively. The dataset of734 chemicals was divided into a training set ( =675), a vahdation set (n=38) and a test set (n=21). A comparison of the MLR results (training, r =0.94, s=0.58 vahdation r =0.84, s=0.67 test, r =0.80, s=0.87) and the ANN results (training, r =0.96, s=0.51 vahdation r =0.85, s=0.62 tesL r =0.84, s=0.75) indicates a smah improvement for the neural network model with five hidden neurons. These QSPR models may be used for a fast and rehable computahon of the aqueous solubihty for diverse orgarhc compounds. 

Use of Broad Biological Profiling as a Relevant Descriptor to Describe and Differentiate Compounds Structure-/ Vitro (Pharmacology-ADME)-/n Vivo (Safety) Relationships... 

The model was claimed to compute 5000-6000 molecules per min. The predictive ability of the model was validated by four approaches. In the first approach, a set of 20 compounds was randomly selected as an initial validation test set. A model was developed from the remaining 86 compounds with an MAE of 0.33, from which the test set values were then predicted. The results of this test prediction were very good and provided momentum for support of the three structure descriptors. In the second approach, a full cross-validation test of the model was investigated. The data set of 102 compounds was divided... 

In an effort to understand if HAMs are, in any way, different from LAMs (low activity molecules), we extended this survey to compounds published between 1991 and 2002, as indexed in WOMBAT [26]. This database [27] contains 4927 unique structures with at least one measured activity better than 1 nM (HAMs), and 34028 unique structures with at least one activity less than 1 XM (LAMs). Between HAMs and LAMs, 1080 molecules are common, that is, they have at least one activity above 1 nM and at least one activity below 1 XM. This is not uncommon for, for example, highly selective molecules. We did not exclude these from either set since we monitor trends, not exact figures. We studied these trends using 2-D-(two-dimensional) descriptors, that is, descriptors that do not use information related to the three-dimensional characteristics of model compounds. These descriptors can be classified as follows ... 

Martin s much-cited comparison of clustering methods and structural descriptors for compound selection (28). 

Junghans, M. and Pretsch, E. (1997) Estimation of partition coefficients of organic compounds local database modeling with uniform-length structure descriptors. Fresenius Journal of Analytical Chemistry, 359, 88-92. 

---