Chemical pattern recognition techniques

Starting from the pioneering work of Kowalskii published in the year of 1972, various chemical pattern recognition techniques have been the powerful tools in SAR work. 

Metabolomics studies the entire metabolism of an organism. It is possible to consider characterising the complex pattern of cellular proteins and metabolites that are excreted in urine. Pattern recognition techniques of nuclear magnetic resonance spectra have been applied to determine the dose-response using certain classical liver and kidney toxicants (Robertson et al, 2000). This could well provide a signature of the functional state of the kidney, and perturbations in the pattern as a result of exposure to a chemical could be observed. But first it would be necessary to understand how compounds with known effects on the kidney affect these processes. 

While principal components models are used mostly in an unsupervised or exploratory mode, models based on canonical variates are often applied in a supervisory way for the prediction of biological activities from chemical, physicochemical or other biological parameters. In this section we discuss briefly the methods of linear discriminant analysis (LDA) and canonical correlation analysis (CCA). Although there has been an early awareness of these methods in QSAR [7,50], they have not been widely accepted. More recently they have been superseded by the successful introduction of partial least squares analysis (PLS) in QSAR. Nevertheless, the early pattern recognition techniques have prepared the minds for the introduction of modem chemometric approaches. 

Several statistical and pattern recognition techniques were used to unravel the relationships between chemical reactivity data and the previously described effects which influence them. 

Data have been collected since 1970 on the prevalence and levels of various chemicals in human adipose (fat) tissue. These data are stored on a mainframe computer and have undergone routine quality assurance/quality control checks using univariate statistical methods. Upon completion of the development of a new analysis file, multivariate statistical techniques are applied to the data. The purpose of this analysis is to determine the utility of pattern recognition techniques in assessing the quality of the data and its ability to assist in their interpretation. 

For example, a single estimate for total PCB s has been historically collected in the NHATS program. Current advances in chemical analysis protocols now allow for the determination of isomer specific resolution of PCB s. Given the 209 PCB s that are now possible to detect, an adequate evaluation of the data without the use of pattern recognition techniques seems impossible. From a QA/QC perspective, these methods can facilitate the detection of outliers and aid in the interpretation of human chemical residue data. The application of statistical analysis must keep abreast with these advances made in chemisty. To handle the complexity and quantity of such data, the use of more sophisticated statistical analyses is needed. 

Two pattern recognition techniques are applied to the analysis of the library of FTIR spectra compiled by the US EFA> The patterns which emerge demonstrate the influence of molecular structure on the spectra in a way familiar to chemical spectroscopists They are also useful in evaluation of the library, which is not error free, and in assessing the difficulties to be expected when using FTIR spectra for complex mixture analysis. 

The main advantage of piezoelectric devices is that, in principle, any process that results in a mass change at an interface can be measured. However, this very nonselective transduction process is also a major disadvantage in that it mandates the use of even more selective surface chemistries than are required for other types of chemical transducer systems. This will make the implementation of piezoelectric chemical sensing devices for ocean measurements rather difficult, but by no means impossible. Indeed, the coupling of pattern recognition techniques with an array of marginally selective piezoelectric transducers may, in the future, make these devices more useful for quantitative ocean measurements. 

Recent developments in the field of sensing airborne chemicals using electrochemical sensors and sensor arrays are reviewed. Such systems detect, Identify, and quantify potential chemical hazards to protect the health and safety of workers and citizens. The application discussed In this review article Is single chemicals at part-per-million levels in air. The sensor system consists of an array of sensors used In four modes of operation, and the data are Interpreted by a computer algorithm. Pattern recognition techniques are being used to understand the information content of the arrays and to focus future experimental work. 

Subsequently, other researchers developed the electronic nose idea with a variety of chemical gas sensor arrays using different pattern recognition techniques for improving the interpretation of responses [2-5]. 

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. 

Figure I. Flow chart of steps involved in structure-activity studies using chemical structure information handling and pattern recognition techniques...

Progress in chemometrics has made a number of new statistical techniques available, which are increasingly being used. This concerns both new supervised and unsupervised (or pattern recognition ) techniques. Chemometrics was dehned about 25 years ago as the chemical discipline which uses mathematical, statistical and related techniques to design optimal measurement procedures and experiments, and to extract maximum relevant information from chemical data. The science of chemometrics has been developed to promote applications of statistics in analytical, organic and medicinal chemistry. 

Copies of the TNO peroxide test databases have been provided to E27.07 and the new versions of CHETAH are expected to contain an extensive database as well as pattern-recognition techniques for estimating the hazard of new materials. The CHETAH software will continue to rely on bond energy data and group contribution calculations to estimate energy release potential. Hopefully, the new versions will also incorporate natural language expert system-type front ends so that the CHETAH program(s) will see expanded use in both analytical and tutorial modes. Copies of the LEILA (8) dissertation have also been provided to E27.07 as an example of an expert system approach to selection and use of appropriate theories and computational methods for the solution of problems in chemical kinetics. 

Our own efforts to miniaturize these instruments are driven by two major interests biological agent detection and clinical diagnostics. The development of a biological threat sensor (rather than a chemical sensor) has been addressed in the past by instruments using pattern recognition techniques for spectra... 

Part of Chapter 11 and Chapter 12 explore the use of, firstly, pattern recognition techniques to enhance the performance of the chemical sensors and secondly the modulation of the smell intensity to improve the signal processing. 

For all those above reasons, trying to classify environmental odour sources in real life with pattern recognition techniques gives rise to a spread of observation points in the different clusters. For the signal processing, that implies to have a great number of samples in order to consider the various conditions and chemical compositions. But that induces a high dispersion of the data of a same class and/or of the same concentration. 

Computer-Assisted Studies of Chemical Structure and Olfactory Quality Using Pattern Recognition Techniques... 

---