Biological activity prediction

Expert systems have also been devised for predicting biological activity. Predicting biological activity is discussed further in Chapter 38. 

A Brief Review of the QSAR Technique. Most of the 2D QSAR methods employ graph theoretic indices to characterize molecular structures, which have been extensively studied by Radic, Kier, and Hall [see 23]. Although these structural indices represent different aspects of the molecular structures, their physicochemical meaning is unclear. The successful applications of these topological indices combined with MLR analysis have been summarized recently. Similarly, the ADAPT system employs topological indices as well as other structural parameters (e.g., steric and quantum mechanical parameters) coupled with MLR method for QSAR analysis [24]. It has been extensively applied to QSAR/QSPR studies in analytical chemistry, toxicity analysis, and other biological activity prediction. On the other hand, parameters derived from various experiments through chemometric methods have also been used in the study of peptide QSAR, where partial least-squares (PLS) analysis has been employed [25]. 

In this chapter we overview some probabilistic methods used for biological activity prediction, paying particular attention to the problems of creation of the training and evaluation sets, validation of (Q)SAR models, estimation of prediction accuracy, interpretation of the prediction results and their application in virtual screening. 

Thus, probabilistic biological activity prediction methods can be used for both estimation of adverse/toxic effects in molecules under study and for finding the multi-targeted ligands, which might yield drugs of superior clinical value compared with monotargeted formulations ... 

Leave one out cross-validation for 3300 kinds of biological activity and 117 332 substances provides the estimate of PASS prediction accuracy during the training procedure. The average accuracy of prediction is about 94.7% according to the LOO CV estimation, while that for particular kinds of activity varies from 65% [System lupus erythematosus treatment, Immunomodulator (HIV)] to 99.9% (Allergic rhinitis treatment, histone acetylation inducer). The estimated accuracy of prediction for all kinds of biological activity predicted by PASS is presented at the web site. " ... 

To identify potentially active compounds in the virtual library, FOCUS-2D employs stochastic optimization methods such as SA (228, 229) and (jA (230-232). The latter algorithm was used for targeted pentapeptide library design as follows. Initially, a population of 100 peptides is randomly generated and encoded by use of topological indices or amino acid-dependent physicochemical descriptors. The fitness of each peptide is evaluated by its biological activity predicted from a precon-structed QSAR equation (see below). Two par-... 

The planned use of QSAR model predictions is an important factor to take into consideration in physico-chemical properties and biological activities prediction, and in virtual screening aimed at prioritizing and planning the design of safer alternatives. The primary focus, also in regulation, should be predictive ability verified on new chemicals, while descriptor interpretations are secondary. The order of OECD principles must be followed. A preconceived notion of what descriptors mean can be a potential source of error in SAR interpretation. Even a minute change in the compound structure can result in a substantial activity... 

Many of the products are biologically active. Predictive models of the toxicity of products, intermediates and by-products in the pure state and as mixtures would be useful at an early stage, and may help to decide between competing process options. 

The second application of the CFTI protocol is the evaluation of the free energy differences between four states of the linear form of the opioid peptide DPDPE in solution. Our primary result is the determination of the free energy differences between the representative stable structures j3c and Pe and the cyclic-like conformer Cyc of linear DPDPE in aqueous solution. These free energy differences, 4.0 kcal/mol between pc and Cyc, and 6.3 kcal/mol between pE and Cyc, reflect the cost of pre-organizing the linear peptide into a conformation conducive for disulfide bond formation. Such a conformational change is a pre-requisite for the chemical reaction of S-S bond formation to proceed. The predicted low population of the cyclic-like structure, which is presumably the biologically active conformer, agrees qualitatively with observed lower potency and different receptor specificity of the linear form relative to the cyclic peptide. 

The protein folding problem is the task of understanding and predicting how the information coded in the amino acid sequence of proteins at the time of their formation translates into the 3-dimensional structure of the biologically active protein. A thorough recent survey of the problems involved from a mathematical point of view is given by Neumaier [22]. 

In chemoinformatics, chirality is taken into account by many structural representation schemes, in order that a specific enantiomer can be imambiguously specified. A challenging task is the automatic detection of chirality in a molecular structure, which was solved for the case of chiral atoms, but not for chirality arising from other stereogenic units. Beyond labeling, quantitative descriptors of molecular chirahty are required for the prediction of chiral properties such as biological activity or enantioselectivity in chemical reactions) from the molecular structure. These descriptors, and how chemoinformatics can be used to automatically detect, specify, and represent molecular chirality, are described in more detail in Chapter 8. 

The most important task of modeling is prediction. The model itself is needed for evaluating the biological activities (and/or physical properties) of compounds, where it is either difficult or costly to measure the activities experimentally. 

The HYBOT descriptors were successfully applied to the prediction of the partition coefficient log P (>i--octanol/water) for small organic componnds with one acceptor group from their calculated polarizabilities and the free energy acceptor factor C, as well as properties like solubility log S, the permeability of drugs (Caco-2, human skin), and for the modeling of biological activities. 

The possibilities for the application for neural networks in chemistry arc huge [10. They can be used for various tasks for the classification of structures or reactions, for establishing spcctra-strncturc correlations, for modeling and predicting biological activities, or to map the electrostatic potential on molecular surfaces. 

GAs or other methods from evolutionary computation are applied in various fields of chemistry Its tasks include the geometry optimization of conformations of small molecules, the elaboration of models for the prediction of properties or biological activities, the design of molecules de novo, the analysis of the interaction of proteins and their ligands, or the selection of descriptors [18]. The last application is explained briefly in Section 9.7.6. 

---