Extraction of information from large data sets by pattern recognition

The use of computers has made data collection much easier and analytical chemists increasingly wonder how to make use of all the data obtained. Pattern recognition permits to extract information present in large data sets in an automatic way.Many scientists acknowledge this fact but are rebutted by the task of learning to use pattern recognition methods. Indeed, there are many methods available and for the newcomer it is extremely difficult to make a selection. For this reason, the lecture will start by explaining the models used in pattern recognition. This will be followed by a critical discussion of advantages and disadvantages of the methods and a selection of preferred methods.

---

Gewinnung von Information aus großen Datenmengen mit Hilfe der Strukturerkennung

     

Kinetic Methods of Analysis: How Do They Rate?

During the past few years there has been a resurgence of interest in kinetic aspects of analytical chemistry and in kinetic methods of analysis. The increased activity is a result of advances that have been made in instrumentation and in data processing techniques. Still, the kinetic approach is not widely applied by practicing analytical chemists. Why is this true? Why are kinetic methods not among the most common methods in use by analytical chemistry? This paper discusses the advantages and limitations of kinetic methods and addresses the probable reasons that they are not widely used. Several new principles have cmerged that are guiding the development of new kinetic-based determinations. These developments have made it possible to compensate for errors that result from changes in reaction conditions and to determine accurately multiple components in mixtures. With these advances kinetic methods are approaching the reliability of traditional equilibrium-based determinations and should be re-evaluated by analytical chemists.     

Chemistry in Times of Artificial Intelligence

Chemists have to a large extent gained their knowledge by doing experiments and thus gather data. By putting various data together and then analyzing them, chemists have fostered their understanding of chemistry. Since the 1960s, computer methods have been developed to perform this process from data to information to knowledge. Simultaneously, methods were developed for assisting chemists in solving their fundamental questions such as the prediction of chemical, physical, or biological properties, the design of organic syntheses, and the elucidation of the structure of molecules. This eventually led to a discipline of its own: chemoinformatics. Chemoinformatics has found important applications in the fields of drug discovery, analytical chemistry, organic chemistry, agrichemical research, food science, regulatory science, material science, and process control. From its inception, chemoinformatics has utilized methods from artificial intelligence, an approach that has recently gained more momentum.     

Quality and the development of reference materials, including the role of traceability and comparability

The global recognition that quality is an economic issue is requiring analytical chemists to look at the chemical measurement process in a way that has not been done before. Much work has been done in certifying reference materials, writing measurement protocols, creating measurement networks, developing analytical measurement techniques and other efforts to make good measurements. This article explores the meaning of quality in chemical measurements and discusses quality in terms of credibility, reliability, traceability and comparability. The importance of understanding the contribution of comparability and traceability to quality in chemical measurements and chemical metrology is emphasized.     

Open-access liquid chromatography/mass spectrometry in a drug discovery environment

The use of open-access mass spectrometry to monitor synthetic chemistry reactions, and also the integrity and purity of new chemical entities, has been a part of the medicinal chemist's tool-box for more than 5 years. Originally in our group at Wyeth Research there were two open-access methods available to the chemists, flow injection analysis (FIA) and liquid chromatography/mass spectrometry (LC/MS). The FIA method was approximately 3 min long, while the LC/MS method was approximately 20 min long (including an 8 min gradient). Within the first 2 years, the total number of open-access analyses increased by approximately 125%. It is interesting, however, that the number of LC/MS analyses increased by more than 285%. This is attributed to the fact that the chemists began using the LC/MS data to monitor reactions and also to check final product integrity and purity. In addition, the number of chemists performing parallel synthesis reactions has increased; thus, individual chemists can produce sample sets of up to 100 vials. This paper describes the implementation of new methodology, which accommodates the need for much faster run times and also the ability to acquire alternating positive and negative ion spectra within the same run. In addition, the instrument has been configured to e-mail the resulting processed data report to the submitting chemist. Several methods have been developed, including structure elucidation using in-source collision-induced dissociation (CID) and night-time analysis. The LC/MS methods for this system are described herein and are applicable to both industrial and academic synthetic chemistry optimization efforts.     

Anwendung einiger Methoden der automatischen Zeichenerkennung auf die interpretation von Massenspektren

Some pattern recognition methods are briefly discussed from the geometric point of view: classification by distance measurements to prototypes, classification by decision-planes and classification by the nearest-neighbor-method. The last two methods give good results with regard to the automatic determination of molecular structures from low resolution mass spectra. It is emphasized, that pattern recognition methods may be useful for interpreting various types of data pools in physical and analytical chemistry.     

Non-selective chemical sensors in analytical chemistry: from “electronic nose” to “electronic tongue”

Development, recent historical background and analytical applications of promising sensor instruments based on sensor arrays with data processing by pattern recognition methods have been described. Attention is paid to the “electronic tongue” based on an array of original non-specific (non-selective) potentiometric chemical sensors. Application results for integral qualitative analysis of beverages and for quantitative analysis of biological liquids and solutions, containing heavy metals are reported. Discriminating abilities and precision obtained allow to consider “electronic tongue” as a perspective analytical tool.     

Solving chemical problems of environmental importance using cavity ring-down spectroscopy

Cavity ring-down (CRD) is a sensitive variant of traditional absorption spectroscopy that has found increasing use in a number of chemical measurement applications. This review focuses on applications of cavity ring-down spectroscopy that will be of interest to environmental chemists and analytical chemists working on environmental problems. The applications are classified into direct monitoring approaches, indirect analysis methods and ancillary studies and a differentiation is made between field-tested instruments and proof of principle studies.     

Determination of mycotoxins in human foods

This tutorial review deals with the analytical methods available for the determination of mycotoxins in food commodities. As the secondary metabolites of a range of fungal species, mycotoxins possess diverse chemical structures, presenting analytical chemists with a unique set of challenges in the microg kg(-1) (ppb) range. A number of analytical methods have been applied to mycotoxin analysis. These include widely applicable HPLC methods with UV or fluorimetric detection, which are extensively used both in research and for legal enforcement of food safety legislation and for regulations in international agricultural trade. Other chromatographic methods, such as TLC and GC, are also employed for the determination of mycotoxins, whereas recent advances in analytical instrumentation have highlighted the potential of LC-MS methods, especially for multi-toxin determination and for confirmation purposes. Conventional chromatographic methods are generally time consuming and capital intensive, and hence a range of methods, mostly based on immunological principles, have been developed and commercialised for rapid analysis. These methods include, among others, enzyme-linked immunosorbent analysis (ELISA), direct fluorimetry, fluorescence polarization, and various biosensors and strip methods.     

分析化学计量学

本文是《分析试验室》第五篇“分析化学计量学”专题定期评述文章。评述了１９９７年１月至１９９８年１２月我国分析化学计量学的主要进展。内容涉及统计学与统计方法、试验设计与优化、分析信号处理、多元校正、化学模式识别、定量构效关系（ＱＳＡＲ）、数据库及专家系统、化学计量学教学等方面,共引用文献１６８篇。     

Non-selective chemical sensors in analytical chemistry: from “electronic nose” to “electronic tongue”

Development, recent historical background and analytical applications of promising sensor instruments based on sensor arrays with data processing by pattern recognition methods have been described. Attention is paid to the “electronic tongue” based on an array of original non-specific (non-selective) potentiometric chemical sensors. Application results for integral qualitative analysis of beverages and for quantitative analysis of biological liquids and solutions, containing heavy metals are reported. Discriminating abilities and precision obtained allow to consider “electronic tongue” as a perspective analytical tool. Received: 17 July 1997 / Revised: 19 February 1998 / Accepted: 24 February 1998     

Identification of different types of imperial age marble finds using instrumental chemical analysis and pattern recognition analysis

A physical-chemical characterisation of several marbles frequently used in ancient times for artistic or decorative purposes was performed in support the work of historians and restorers. The data were obtained using several different types of instrumental chemical methods (Thermogravimetry, Differential Thermal Analysis, X-ray Diffractometry and ICP Plasma Emission Spectroscopy) and have been summarised in short tables. The data have already proved useful in the identification of a small number of finds (statues or architectonic elements) from Ancient Rome (Imperial Age, 2nd-3nd cent. A.D.) for the purpose of which also a well-known pattern recognition analysis software package was used for data processing. In practice, the research showed that an organised set of chemical data obtained using several modern instrumental methods can provide a valid basis for the reasonably rapid and reliable identification of the type of marble used to make artistic artifacts that have not yet been subjected to typological study.     

Measuring splitting and overlapping in molecular systems. A fuzzy set estimation of reaction hazards

An algorithm for assessments of fuzzy molecular structural characteristics has been presented, and its chemical relevance is approved for numerous evidences of reaction mechanisms. Empirical rules pervading the results of chemical synthesis are continuously the source of information efficiently used by experimental chemists and technologists. Approaching them theoretically to broaden areas of applicability and to make them more precise for better prediction is of great importance to reach progress in the chemical information recognition and to steer and control technological processes. Fuzzy sets called splitting and overlapping have been applied for assessments of the reaction hazards. The mathematically grounded properties turned out to underlie well-known empirical rules for preliminary estimations of organic reaction tendencies. Informative quantitative data shown are revealing new ways of understanding basic chemical reaction mechanisms.     

Higher-throughput screening for Caco-2 permeability utilizing a multiple sprayer liquid chromatography/tandem mass spectrometry system

In the current drug discovery environment, higher-throughput analytical assays have become essential to keep pace with the screening demands for drug metabolism and pharmacokinetics (DMPK) attributes. This has been dictated by advances primarily in chemical procedures, notably combinatorial and parallel syntheses, which has resulted in many-fold increases in the number of compounds requiring DMPK evaluation. Because of its speed and specificity, liquid chromatography/tandem mass spectrometry (LC/MS/MS) has become the dominant technology for sample analysis in the DMPK screening assays. For higher-throughput assays, analytical speed as well as other factors such as method development, data processing, quality control, and report generation, must be optimized. The four-way multiplexed electrospray interface (MUX), which allows for the analysis of four LC eluents simultaneously, has been adopted to maximize the rate of sample introduction into the mass spectrometer. Generic fast-gradient HPLC methods that are suitable for approximately 80% of the new chemical entities encountered have been developed. In-house-written software programs have been used to streamline information flow within the system, and for quality control by automatically identifying analytical anomalies. By integrating these components together with automated method development and data processing, a system capable of screening 100 compounds per week for Caco-2 permeability has been established.     

Neighbourhood Sum Degree-Based Indices and Entropy Measures for Certain Family of Graphene Molecules

A topological index (TI) is a real number that defines the relationship between a chemical structure and its properties and remains invariant under graph isomorphism. TIs defined for chemical structures are capable of predicting physical properties, chemical reactivity and biological activity. Several kinds of TIs have been defined and studied for different molecular structures. Graphene is the thinnest material known to man and is also extremely strong while being a good conductor of heat and electricity. With such unique features, graphene and its derivatives have found commercial uses and have also fascinated theoretical chemists. In this article, the neighbourhood sum degree-based M-polynomial and entropy measures have been computed for graphene, graphyne and graphdiyne structures. The proper analytical expressions for these indices are derived. The obtained results will enable theoretical chemists to study these exciting structures further from a structural perspective.     

Electrophoresis of cereal storage proteins

Cereal proteins have been studied by a number of analytical techniques over the years. One of the major methodologies utilized by cereal chemists has been electrophoresis. Starting with moving boundary electrophoresis and progressing to slab gels and high-performance capillary electrophoresis, innovative methods have been developed to provide high resolution separations of difficult to separate proteins. Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), acid-PAGE, isoelectric focusing, free zone CE, and even high-resolution two-dimensional HPLC-HPCE methods have been developed to separate cereal proteins. This review focuses on electrophoretic methods for separating and characterizing cereal storage proteins.     

分析化学计量学

本文是《分析试验室》定期评述中“分析化学计量学”课题的第一篇评述文章,收集了1980年至1990年国内有关期刊和会议资料,对分析化学计量学在我国近10年来的发展概况作了评述,内容包括基础理论、试验、优化设计与优化方法、分析校正、分析信号处理、因子分析、化学模式识别、结构解析、库检索和计算机应用等。     

Chemometrics and food chemistry: data validation

Chemometrics has often been applied in food chemistry to cluster and classify samples or to produce models for food quality. In recent years, data on food composition have become important for public health protection and food trades. The quality of the available chemical information on foods is a problem; data obtained with the newer analytical methods is scarce and general knowledge about food composition is poor, judged by published tables on food composition. Moreover, agreement between results obtained by different analytical methods is very poor. To overcome this critical problem, several countries have decided to create data banks on food composition. The analytical results to be stored must be validated. Chemometric modelling is useful for this purpose. Interlaboratory studies allow standardization of methods and the preparation of food reference materials. The classical computation of repeatability and reproducibility does not extract all the available information so that a multivariate approach is necessary to improve the quality of a data bank on food composition.     

Improving the analytical performances of inductively coupled plasma optical emission spectrometry by multivariate analysis techniques

The various multivariate analysis techniques which have been successfully applied to maximize the analytical performance of ICP-OES are reviewed. These include optimization procedures, spectral data processing and calibration methods as well as classification and pattern recognition techniques.