Hidetsugu Abe

Extended studies of the automated odor-sensing system based on plural semiconductor gas sensors with computerized pattern recognition techniques

Abstract The odor-sensing system reported earlier is used to measure odors of 47 compounds. Three computerized pattern recognition techniques, the k -NN, simplex and potential function methods, are applied to the data set for the prediction of the odors of the corresponding compounds. The best prediction results (80.9%) were obtained by means of the potential function method.

Automated odor-sensing system based on plural semiconductor gas sensors and computerized pattern recognition techniques

Abstract Although odor is an inherent characteristic of chemical substances, no methods to measure or identify odor objectively have been available. A computerized method for classification and identification of odor is reported. Odors are detected by using eight semiconductor gas-sensor elements which have different sensing properties for gases; the output from the odor-detecting apparatus for a gas sample is represented by an 8-dimensional vector. The vectors for 30 substances were examined by clustering analysis and four obvious clusters were observed. These clusters corresponded to ethereal, etherealminty, ethereal-pungent and pungent substances.

Computer-aided structure elucidation of organic compounds with the chemics system: Removal of Redundant Candidates by 13C-n.m.r. Prediction

Abstract A 13 C-n.m.r. prediction module capable of removing inappropriate candidate structures given for an unknown compound based on the spectral data is introduced for the CHEMICS system. Given a set of candidate structures generated in the system, the routine may be used to prune off redundant candidates which have a predicted number of signals inconsistent with the observed number. It is shown that the addition of the examination module to the system makes structure elucidation by computer much more practical.

Systemization of semantic descriptions of odors

Abstract Semantic description is the only practical method of representing the quality of odors but it can be very subjective. In an attempt to improve the objectivity, a set of 126 odor descriptors corresponding to 1573 organic compounds was studied. Special overlap coefficients are calculated to express similarities between descriptors and the breadth of meaning of some terms. Cluster analysis shows that there are 19 categories of odor. These categories agree with earlier proposals for classification of primary odors.

A structure-biological activity study based on cluster analysis and the nonlinear mapping method of pattern recognition

Abstract Cluster analysis is used in a study of structure—activity relationships of biologically active compounds. A hierarchal clustering technique was applied to 29 typical antibiotics using 27 antibacterial activities. These antibiotics were of various types; penicillins, cephalosporins, aminoglycosides, macrolides, tetracyclines, and peptides. The result was obtained as a branching tree diagram. The technique allowed the antibiotics to be distributed into 6 clusters, each cluster mostly consisting of compounds with a similar structure. Nonlinear mapping was used to display the 27-dimensional data structure of the antibiotics. The nonlinear map was compared with the clusters obtained by cluster analysis.

Microcolumn Liquid Chromatography Combined with Computer-Assisted Retention Prediction System for Polycyclic Aromatic Hydrocarbons in Extract from Diesel Particulate Matter

Abstract Recent advances in column and instrument technology have made the development of d new generation of high-resolution microcolumn liquid chromatography possible. In addition to offering reduced solvent consumption, this chromatographic technique also yields higher mass sensitivities than those in conventional systems. In this study, the applicability of this technique to the analysis of polycyclic aromatic hydrocarbons (PAH) is investigated. PAH in the extract from diesel particulate matter were analyzed to demonstrate the utility of this approach combined with the computer-assisted retention prediction. The technique proposed in this study makes very clean and high cost-performance environmental analysis possible.

Generation of stereoisomeric structures using topological information alone

An algorithm for enumeration of stereoisomeric structures due to asymmetric carbon, C-C double bond and so on has been developed. By using this algorithm, all the possible stereochemical structures for a molecule may be generated on the basis of its topological representation. The identification of each distinct stereoisomeric structure is performed by a modification of Morgans method.

Computer-assisted structure—carcinogenicity studies on polycyclic aromatic hydrocarbons by pattern recognition methods

Abstract Pattern recognition methods are applied to the study of structure—carcinogenicity relationships in 25 representative polycyclic aromatic hydrocarbons (PAHs). On the basis of presumed metabolic transformation, a variety of reactivity indices taken from simple Huckel molecular orbital theory for not only parent PAH but also later metabolites are used to investigate the carcinogenic process. In order to display the 12-dimensional molecular descriptor space, a Karhunen—Loeve plot in two-dimensional space is employed, 92.1% of the variance is retained. The data structure shows asymmetric character. Carcinogens are clustered, whereas non-carcinogens are scattered. Linear discriminant functions for carcinogenicity are developed by using multiple linear regression analysis. The most significant equations suggest the importance of metabolic pathways.

A convenient notation system for organic structure on the basis of connectivity stack

A convenient notation system for organic structures has been developed for the application of the connectivity stack. A notation arbitrarily encoded for a structure by a user through a rather simple procedure using 35 codes, which have been previously prepared, is automatically canonicalized in a computer. The notation given by the user is standardized according to the rules for rearranging the codes into a dictionary order. The connectivity stack is estimated for each of the standard notations and its permuted derivatives. The notation whose stack is the largest amount is decided to be canonical. This notation method will be widely applicable in the field of structure manipulation because of its extreme simplicity.

Computer-assisted structure—carcinogenicity studies on polynuclear aromatic hydrocarbons by pattern recognition methods: The role of the bay and l-regions

Abstract Computer-assisted studies of structure—carcinogenicity relations for unsubstituted polynuclear aromatic hydrocarbons (PAHs) are described. On the basis of the bay and the K- and L-region theories, the carcinogenic process for each PAH is expressed by multi-dimensional descriptors. Factor analysis is used to group these descriptors. Descriptors grouped by the method were found to be useful in understanding the physicochemical properties related to the carcinogenic process, and both the L-region and the bay region were shown to play important roles in the explanation of the carcinogenicity of PAHs.