Pattern recognition chemical structure—biological

Studies of Chemical Structure-Biological Activity Relations Using Pattern Recognition... 

Jurs, P. C., Chou, J. T. and Yuan, M. (1979). Studies of chemical structure-biological activity relations using pattern recognition. In Computer-Assisted Drug Design. 

Catalyst [67] was launched 1992 by BioCAD (now Accelrys) as a tool for automated pharmacophore pattern recognition in a collection of compounds based on chemical features correlated with three-dimensional structure and biological activity data. 

Applications of pattern recognition methodology to chemical problems were first reported in the 1960 s (20,21) with studies of mass spectra. Since then papers have described work in a variety of areas (22,23) including mass spectrometry, infrared spectroscopy, NMR spectroscopy, electrochemistry, materials science and mixture analysis, and the modeling of chemical experiments. Diagnosis of pathological conditions from sets of measurements made on complex biological mixtures, e.g., serum, have been reported (24). The successes in these areas have led to the belief that these methods should prove useful in the development of structure-activity relations. 

It is obviously not possible to unravel the entire complexity of the physical, chemical and biological properties of even the simplest of molecules. However, focusing on the apparently pertinent descriptors for structures, one can, via pattern recognition, begin to equate toxicological response with structures ... 

The spectrum of a compound represents the molecular structure in the form of a complex code. Therefore, a relationship may be expected between spectrum and biological activity. Such a spectra-activity relationship assumes that the mechanism of biological activity is similar to the physical and chemical processes which produce spectra. This similarity is at least doubtful and the direct way of structure-activity relationships seems to be more promising. An attempt to correlate mass spectra with biological activity has led to violent replies in the literature. Controversies on this theme made many chemists very suspicious against applications of pattern recognition. 

Three different approaches have been used for computer-assisted interpretations of chemical data. 1. Heuristic methods try to formulate computer programs working in a similar way as a chemist would solve the problem. 2. Retrieval methods have been successfully used for library search (an unknown spectrum is compared with a spectral library). 3. Pattern recognition methods are especially useful for the classification of objects (substances, materials) into discrete classes on the basis of measured features. A set of characteristic features (e.g. a spectrum) of an object is considered as an abstract pattern that contains information about a not directly measurable property (e.g. molecular structure or biological activity) of the object. Pure pattern recognition methods try to find relationships between the pattern and the "obscure property" without using chemical knowledge or chemical prejudices. 

These complementarity rules owe their discovery to the chemical analysis of DNA by Chargaff and associates (3). The DNA from many different organisms shows the same patterns of base composition, namely A and T are always present in equal quantities, as are G and C. The immediate corollary of this observation, that a purine base (R) exists for every pyrimidine base (Y) and vice versa, led Watson and Crick to propose that two helical strands in DNA are held together by specific, intermolecular purine-pyrimidine (R Y) interactions (4). In turn, this unique chemical complementarity of the double-helical structure, proved to be a major breakthrough to understand the self-recognition and self-reproduction of DNA and forms the cornerstone of structmal biology as we know it today, more than half a century later. 

---