Knowledge-based prediction computational models

In Chapter 4, David Lewis introduces computer-assisted methods in the evaluation of chemical toxicology. He points out that any substance can be toxic, and thus it is the dose of the substance that determines a toxic response. How, then, does one predict toxicity Lewis examines QSAR methods, pattern recognition techniques, computer modeling, and knowledge-based systems to answer this question. Ideally, one would like to assess toxicity of a structure before the compound is synthesized. To bring all this into focus, emphasis is placed on the cytochromes P450. 

As described elsewhere in this book, SARs of increasing sophistication have been used for some time to make toxicity predictions for a variety of organ systems, including the developing embryo/fetus.20 In silico or computer-based analyses are of two primary kinds knowledge or "rule-based" systems and correlative or "statistically based" systems. The details of the different prediction systems have been reviewed in detail elsewhere,21-22 but for the purposes of developmental toxicity prediction two models are discussed briefly here. 

Despite its importance, our knowledge about GO is still limited [27, 28], For example, due to its amorphous and nonstoichiometric character, the atomic structure of GO is not very clear. Many experimental studies on GO structure have been reported. However, structure characterization of systems as complicated as GO requires a combination of both experimental and theoretical efforts. Recently, there are significant progresses on the theoretical sides. First-principles calculations have been used to compare stabilities of different GO structure models. More importantly, computational spectroscopy strategy has also been used to make a direct comparison with experimental data and thus obtain deeper insights for GO structure. At the same time, electronic structure and other properties of GO can also be predicted by computational studies. The most difficult part is the understanding of the mechanisms of oxidation and reduction, which also requires an intense theoretical study. In this review, we will focus on recent progresses in theoretical studies on GO. Structure characterization based on computational spectroscopy is specially emphasized. 

The computational approaches to predict protein structures from their amino acid sequences can be either empirical or knowledge-based. The former involves calculations of the energetically favorable structures by the use of the parameterized force fields (section 9.2) and the latter performs statistical analysis with references to the existing structures employing ranges of bioinformatic techniques (section 16.5). The comparative structural analysis leads to homologous models, which are then refined by energy minimization. 

Trelea, I. C., Titica, M., Landaud, S., Latrille, E., Corrieu, G., Cheruyb, A. (2001). Predictive modeling of brewing fermentation from knowledge-based to black-box models. Mathematics and Computers in Simulation, 56, 405-424. 

---