Metabolomics intends holistic analyses of enormous numbers of organic compounds with different physicochemical properties. This constitutes a tremendous challenge for analytical chemistry when nearly infinite numbers of unknown metabolites are simultaneously analyzed. In other words, for successful metabolomics, every analytical platform should be customized to meet different research objectives. Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR/MS) offers extremely high levels of resolution and sensitivity. The high-resolving power allows direct infusion analyses of crude biological samples without chromatographic separation steps and derivatization reactions, and high-throughput metabolic profiling is possible through assignment of only one or a few molecular formulas to each single m/z value in the spectra. With the aid of the extreme resolving power FT-ICR/MS, we have developed a metabolomics platform of high-throughput metabolic profiling and metabolite candidate identification integrating a data processing system, the Dr.DMASS program (http://kanaya.naist.jp/DrDMASS/), and a metabolite-species database, KNApSAcK (http://kanaya.aist-nara.ac.jp/KNApSAcK/). We discuss the potential of this FT-ICR/MS-based metabolomics scheme as a general metabolomics tool through metabolic profiling and differential metabolomics studies. A future direction of FT-ICR/MS-based metabolomics is also discussed.

View full abstract

We have investigated multiple category classification of pharmacological activity classes of chemical compounds from their chemical structures using multi-class support vector machine (SVM). For the input to the SVM, every chemical structure was represented by a multidimensional pattern vector that was obtained by the Topological Fragment Spectra (TFS) method. In this work, we adopted the "one against the rest" method, which combine two-class SVMs to solve the multiclass classification problem. For the computational trial, we employed 98,634 compounds that belong to 100 different activity classes. The data set was divided into three groups: a training set of 59,180 compounds, a validation set of 29,590 compounds, and a test set of 9,864 compounds. The SVM model was trained using the training set and the validation set. For the test set that consists of 100 classes, the best model correctly classified 80.8% of the drugs into their own active classes including the multi-labeled compounds. The resulted classifier would be useful in drug-discovery and also helpful in risk estimation of adverse effects.

View full abstract

Since some of the descriptors for QSAR data show nonlinear relationships between the descriptors and biological activities, they should be analyzed by nonlinear analysis methods. However, the algorithm of nonparametric regression methods is complicated and is difficult to show the contributions of QSAR descriptors. Hence, we proposed a new simple nonparametric regression method, which was named SANORA, before. This method is based on the additive model as well as penalized local regression. In addition, this method has comparatively simple algorithm which means that the contributions of predictive variables can be easily estimated. We would like to introduce this method and propose the solution of the overfitting problem when it is applied to the data set which has many predictive variables. In order to avoid this problem and to obtain results faster, we applied the PLS algorithm to reduce dimensions of the data: The latent variables were used as predictive variables for the SANORA analysis of ACE inhibition activity data. As a result of cross-validation, the predictive performance of the constructed model became better.

View full abstract

The main object of this study is to construct a statistical model for estimating drug-likeness of chemical structures. The drug and non-drug molecules have been taken from the database of CMC (comprehensive medicinal chemistry) and ACD (available chemicals directory), respectively, and used to construct drug-likeness models. We have generated 3D structures of these molecules by CORINA and have optimized them using MMFF94s. Then chemical descriptors were calculated with DRAGON 5.4, and the drug-likeness models were constructed by using generative topographic mapping (GTM) and SOM. As with SOM, GTM is nonlinear mapping method to map a point from the high-dimensional input space to a low-dimensional space. In both 2D maps, drug and non-drug molecules have been roughly separated. For the purpose of comparison of smoothness of mapping between GTM and SOM, we have introduced new criteria, RMS of midpoint (RMSM). RMSM is calculated as RMS of mapping error of all midpoints instead of the original data points. As a result of calculation of RMSM of both mappings, we concluded that GTM is able to give slightly better nonlinear mapping than SOM.

View full abstract

Understanding plant metabolism as an integrated system is essential for metabolic engineering aimed at the effective production of compounds useful to human life. The integrated transcriptome and metabolome analyses lead to the identification of unknown genes and their regulatory networks involved in metabolic pathways of interest. Here we report the discovery of two transcription factors that positively control the biosynthesis of glucosinolates, a group of bioactive secondary products, in Arabidopsis thaliana by an integrated omics approach. Combined transcriptome coexpression analysis of publicly-available condition-independent data, and the condition-specific data identified Myb28 and Myb29 as candidate transcription factor genes specifically involved in the regulation of aliphatic glucosinolate production. Analysis of a knockout mutant demonstrated that Myb28 is a positive regulator for basal-level production of aliphatic glucosinolates. Myb29 presumably plays an accessory function for methyl jasmonate-mediated induction of a set of aliphatic glucosinolate biosynthetic genes. Overexpression of Myb28 in Arabidopsis cultured suspension cells, which don't normally synthesize glucosinolates, resulted in the production of large amounts glucosinolates, suggesting the possibility of efficient industrial production of glucosinolates by manipulation of these transcription factors. A working model for regulation of glucosinolate production involving these genes, renamed Production of Methionine-derived Glucosinolate (PMG) 1 and 2, are postulated.

View full abstract

Homology searches have been used widely to predict functions of genes and proteins when genomes are newly decoded. As a result of decoding of genome sequences from a wide range of species, a large number of proteins whose function cannot be estimated by the homology search of amino acid sequences is progressively accumulated and remains of no use in science and industry. A method to estimate the protein function that does not depend on the sequence homology search is in urgent need. We previously developed a Batch-Learning SOM (BL-SOM) for genome informatics, which does not depend on the order of data input. The present study focused on BL-SOM analyses on di to tetra continuous amino acid frequencies. Concerning the oligopeptide frequencies in the 110,000 proteins which have been classified into 2,853 function-known COGs (clusters of orthologous groups of proteins), BL-SOMs that faithfully reproduced the COG classifications were obtained. This indicated that proteins, whose functions are presently unknown because of lack of significant homology with function-known proteins, can be related to function-known proteins with the BL-SOM.

View full abstract

In general, metabolome analysis technology is utilized for comprehensive analysis of ionic metabolites derived from a biological organism. When this technology is applied to industry such as the fields of pharmaceuticals, foods, chemistry, and agriculture, we must handle the quantitative measurement of a large variety of small-molecule organic compounds. Since HPLC is used for separation and analysis of a wide variety of organic compounds, we aimed at the development of metabolome analysis technology using LC-TOF-MS that was the mass spectrometer equipped with HPLC. We chose Waters LCT Premier XE system as LC-TOF-MS and developed an automated analysis system (called Metalys2) of MS spectral data produced from this LC-TOF-MS. In this metabolome analysis, we thoroughly reduce noises and extract compound peak information from the spectral data. Subsequently, Metalys2 enables us to obtain precise mass information of the identified organic compounds. By comparing the obtained precise mass with various databases, we add annotations such as chemical formulas and compound names to each compound. Furthermore, comparison of organic compounds identified by Metalys2 between two samples enables us to elucidate qualitative and quantitative differences in compounds between these samples.

View full abstract

Among chemicals existing in environments, very few are identified as carcinogens. But from the viewpoint of the animal protection, it becomes a social, urgent problem to develop a technology to predict the carcinogenicity of the chemicals from the structures. Up to now, a lot of researches have been done, but it is extremely difficult to predict the carcinogenicity of non-congeners with various chemical structures. The large scale reliable data of the carcinogenicity of non-congener chemicals were collected from IARC, EU, EPA, NTP, ACGIH, and JSOH databases and evaluated into four ranks. For 846 chemicals so collected, 69 descriptors were generated by using Fujitsu CAChe ProjectLeader from their three dimension molecular structures. Relationships between carcinogenic ranks and descriptors were analyzed using support vector machine (SVM) and neural network (ANN) techniques. ANN showed a satisfactory predictability for screening test using animals, but took much computing time. On the other hand, SVM could predict within short time, but its predictability power, inferior to ANN, was considered to be improved by introducing topological descriptors. The prediction of the carcinogenicity of large-scale non-congener chemicals of this research is an extremely complex problem, so the high ability of ANN was proven. On the other hand, SVM has also several advantages out of ANN, so it is expected SVM and ANN are properly used for various nonlinear problems.

View full abstract

Soft sensors are widely used in chemical plants to estimate a process variable which is difficult to measure online. An inferential model is constructed between variables which are easy to measure online and one which is difficult to measure online, and an objective variable is estimated by the model. However, soft sensors have some practical difficulties. One of the crucial difficulties is that predictive accuracy drops due to changes of state of chemical plants, sensor and process drifting and so on. If a problem of degradation of soft sensors is not solved, it is difficult to identify reasons of abnormal situations. There is no effective method to solve these difficulties under the circumstances. In this study, we have developed the new soft sensor method that combines independent component analysis (ICA) and PLS. ICA is a method that is used in many fields such as signal processing. We can comprehend the state of plant by an ICA model and estimate an objective variable by a PLS model, updating PLS model appropriately. We showed the superiority of this method over a traditional one with real industrial data.

View full abstract

We propose two methods for more reliable and accurate property estimation in the field of QSPR. One of those is a method to determine applicability domain (AD) of QSPR model numerically based on pattern recognition of chemical structures. AD is a part of chemical space in which chemical structures are predicted moderately by corresponding QSPR model. A classification model is constructed through supervised learning and its outputs indicate whether a compound is in the AD of a QSPR model or not. On the other hand, we also developed integration method of multiple QSPR models rather than selecting one of them. For each compound, a linear combination of the all outputs of applicable models is formed to minimize error of prediction. Coefficients of linear combination are determined considering model predictivity and correlativity of each pair of models. As a case study, we examined these methods with two data sets of water solubility taken from literatures.

View full abstract

Two-way Boosting Partial Least Squares (TBPLS) regression is described. TBPLS can be seen as an extension of conventional boosting techniques in sample dimensions to variable dimensions (i.e. descriptor). The basic idea of boosting is a series of regressor or classifier obtained by a part of data. In other words, this process can be seen as a kind of resampling based of prediction error by the previous model. In the present study, this idea is extended as boosting in variable dimensions. It proceeds as follows; uninformative variables in the previous are intensively selected for the next model building. This process is repeated several times and final model is obtained as an ensemble of them. Two-way boosting is the simultaneous boosting in both sample and variable dimensions in PLS modeling. This model ensemble technique makes it possible to describe much more latent relationship not only among samples but also variables, such as descriptors. The effect of TBPLS is demonstrated with a QSAR data set.

View full abstract

Partial least squares (PLS) has been widely used in chemometrics studies for the regression and classification of high-dimensional data, especially n (number of samples) << p (number of variables) type data. Canonical correlation analysis (CCA) is a classic method of multivariate analysis. However, it has rarely been applied to multivariate regression in chemometrics studies. The main reason is that, CCA cannot be applied for n<<p type data. So we applied regularized CCA (RCCA), which will add l2-norm penalty to CCA. In this study, we formulated PLS, CCA, RCCA, Kernel PLS and Kernel CCA by using generalized eigenvalue problem. And we applied PLS and CCA to two toy problems to clarify the features of these methods. Finally we applied PLS and RCCA to GC-MS data, which is analyzed to resolve the problems of Japanese green tea ranking. We found that, to construct a regression model, the optimal number of latent variables determined by leave-one-out cross-validation (LOOCV) was significantly fewer in RCCA than in PLS.

View full abstract

Worldwide trend of tighter environmental regulation and prohibition of animal experiments is growing day by day. In silico screening is an ultimate technology to meet this trend. A large problem of in silico screening at present is a low classification and prediction ratio thereof. In the last symposium of Structure-Activity Relationships, the author of this paper reported "KY method" in order to realize a perfect (100%) classification ratio. In this paper, a new method called "Tailor-Made Modeling" will be discussed to solve another problem, that is, improvement of a prediction ratio. This new method significantly improves a prediction ratio as compared to a conventional method. As a result, it is highly expected that in silico screening will be developed to a practical level by the combination of "KY method" and "Tailor-Made Modeling", since the former realizes a perfect (100%) classification while the later significantly improves a prediction ratio.

View full abstract

A novel method for predicting the binding sites for druglike compounds on the surface of proteins was developed on the basis of the specific amino acid composition observed at the ligand-binding sites of ligand-protein complexes determined by X-ray analysis. A profile representing the preference of each of the 20 standard amino acids at the binding sites of druglike molecules was obtained for a small set of high-quality complex structures. An index termed propensity for ligand binding (PLB) was created from these profiles. The PLB index was used to predict the propensity of binding for 804 ligands at all potential binding sites on the proteins whose structures were determined by X-ray analysis. If the sites with the first two highest PLB indices are taken into consideration, the successfully predicted sites reached a high percentage of 86. The PLB prediction is relatively simple, but the validation study showed that it is both fast and accurate to detect ligand-binding sites, especially the binding sites of druglike molecules. The present study has demonstrated that the PLB index is powerful in predicting not only ligand-binding sites of uncharacterized proteins but also novel drug-binding sites of known drug targets.

View full abstract

Identification of the druggable concavity, where druglike molecules are highly inclined to bind, is an important step in structure-based drug design. We previously developed an index named PLB (propensity for ligand binding). The PLB index was proven to be useful in identifying the druggable concavities in the quality X-ray structures of target proteins. Here, we apply the PLB to predicting the druggable concavity in model structures constructed by homology modeling. In this study, we assembled a set of reference proteins that were accurately determined by X-ray analysis in forms of complexes with druglike small molecules. Homology models for the reference proteins were constructed using multiple homologous proteins as templates. The PLB index was then applied to the homology models to predict the druggable concavities. If a template protein in a complex with a druglike small molecule was used, the druggable concavity was predicted well, with a prediction rate of 78%. Interestingly, even when the percent sequence identity between the reference and template proteins was lower than 30, the PLB index could successfully identify the druggable concavity in some cases. This study has demonstrated the practical value of the PLB index in identifying the drugabble concavity in the homology model.

View full abstract

Transition State Database (TSDB) is a system storing information of TSs, reactants, intermediates and products derived from quantum chemical calculations. One of the usages of TSDB is to make initial geometries for new TSs using data in the TSDB and to reduce computational times of searching TSs. Another usage is to confirm whether or not a new route designed by a synthesis route designing system such as AIPHOS, KOSP or TOSP is useful to synthesize target compounds. We have been studying how to use the TSDB for developing new synthesis routes and the data base on the internet. In the present study, we developed a new web interface which search structures in the TSDB. This interface we developed was coded in PHP5 and shows 3D structures, IRC animations and so on. The system is capable of searching chemical structures using the pgchem::tigress program on the PostgreSQL. The program is used to find data in the TSDB by comparing chemical structures of reactants and products. It is possible to make initial geometries by adding substitutions using the data found from the TSDB. Trial version is released at https://trial.tsdb.jp/.

View full abstract

One of the most important characteristics required for ionic liquids (ILs) is the capability to realize low-viscosity. Although ionic liquids are sometimes called "designer solvents", no rational design has yet been proposed for preparing ILs with low viscosity. Among recently developed ionic liquids, we were fascinated by alkyl ether substituted ILs due to their low-viscous nature. We discovered that 1-(3,3'-difluoro)butyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide showed a significant low viscosity compared to the corresponding non-fluorinated version of [bmim][TFSI]. To gain the insight for the rational design of ionic liquids, we attempted to investigate the origin of this change in the viscosity by the difluorination using a DFT calculation. The calculations suggest that there is a clear difference in the population pattern of the stable conformers (probability of occurrence) between bmim cation 1 and 1-(3,3'-difluoro)butyl-3-methylimidazolium cation 3f. Moreover, an intramolecular hydrogen bond is suggested for cation 3f. We will discuss the origin of viscosity change of ILs by the difluorination using DFT calculation method.

View full abstract

Hydrolysis reaction is one of the major sources of degradants in the biodegradation test of Chemical Substance Control Law in Japan. We intended to add a new insight into the biodegradation prediction using activation energies of hydrolysis reaction. By applying Ingold's Aac2 and Bac2 mechanisms of ester hydrolysis, we made calculation of activation energies for the mechanisms using both MOPAC PM3 and DFT calculations and compared the calculated values with experimental data of those esters. And we developed prediction method of hydrolysis activation energy with combined use of MOPAC and DFT. We tried to use residual percentage after 28 days, calculated from predicted activation energies, as a measure of hydrolysis with threshold from water solubility. A resulting residual percentage data was compared to the hydrolysis estimation in the biodegradation test and it was found that the method shows the comparable performance with commercial software. Especially, our method showed successful performance to indicate the esters not to be hydrolyzed.

View full abstract

Fac-tris(2-phenylpyridine)iridium(III)(Ir(ppy)3) is one of the useful luminescent materials for organic EL devices. Recently, it was found that Ir(acac)3 reacts with 2-phenylpyridine (ppyH) to produce Ir(ppy)3 in the presence of Bronsted acids with a high yield. In the present study, we tried to understand the detailed mechanism of the reaction and to explicate the effects of Bronsted acid on the reaction. The B3LYP/SDD level of theory was used to investigate the mechanism of the ligand exchange reaction with and without the acids. A proton addition to an oxygen atom in one of the acac ligand results in breaking an Ir-O bond and changing its structure to a five coordination geometry suitable for accepting an N donor of ppyH. It was confirmed that solvent molecule extracts a proton on the acac ligand before the ligand exchange reaction occurs. Without removing the proton on the ligand, we only obtained a TS producing a by-product in which the elimination of an acac ligand from Ir(acac)3 does not occur. In conclusion, a proton from Bronsted acid and including an H2O molecule are the key to exchanging the acac ligand to the ppyH one in an early stage of the reaction.

View full abstract

Metagenome analyses of mixed genomes of uncultured microorganisms derived from various environmental samples have been progressively developed, and four million or more of gene candidates are registered in the international DNA databanks without information of their gene functions. As a bioinformatics exercise in Nagahama Institute of Bio-Science and Technology, a database for genes that may contribute achievement of the environmentally sustainable world in a broad sense are constructed. Students independently set and search for, on their own motivation, such genes that may contribute environmental cleanup and detoxification from a huge number of gene candidates obtained by metagenome analyses but left of no use because of lack of annotation concerning their functions. Genes newly annotated by students have been registered in the "Database for Genes Contributing Sustainable World" constructed by our group and publicized with students names. Many gene candidates that are thought to have the potentiality in the heavy metal removal, the degradation of agrichemicals, PCB, or environmental hormones, and the bio-fuel production have been found. The exercise text is available by requests.

View full abstract

Promolecules derived from a given skelelton are regarded as RS-stereoisomers and they are categorized into five types (Types I--V) by means of chirality, RS-stereogenicity, and sclerality. They are generally enumerated under the actions of three subgroups (the maximum point subgroup, the maximum RS-permutation group, and the maximum lignd-inversion group). The general method is applied to allene and methane skeletons.

View full abstract

The topological index Z was proposed by the present author in 1971 for characterizing the topological structure of the carbon atom skeleton of hydrocarbon molecules, which was found to have good correlation with the boiling point of alkanes. However, recently Z was found to play a very important role in elementary mathematics and algebraic number theory by transforming algebra into geometry or graph-theory. Although a number of interesting mathematical findings were reported last year in this symposium, further new features were discovered. For example, (i) the Z index of the caterpillar graph is identical to the value of the continuant, if they are composed of the common set of elements, (ii) there are always found a series of caterpillar graphs for a given generalized Fibonacci numbers with the property, Gn=Gn•+Gn•, etc.

View full abstract

We are exclusively utilizing biomolecules, such as enzyme proteins and carbohydrates, in the fields of organic synthesis and bioorganic chemistry. Recently we have focused our attention on computational simulations of lipase enzymes and lipase-ligand complexes to understand and predict the substrate specificity and enantioselectivity of lipases toward non-natural organic compounds on the basis of quantum theory. In this work, Candida antarctica lipase typeB (CALB)-secondary alcohol ester complexes, which were refined by both continuum and explicit solvent models for water, were computed using ab initio fragment molecular orbital (FMO) calculations at MP2/STO-3G level. The all-electron FMO calculations showed that the binding energy for the fast reacting (R)-ester is negatively larger than that for the slow reacting (S)-ester. It is also found that the interaction energies between Thr40 residue in the CALB complexes and the ester substrates are large compared to those between the catalytic triad of CALB, Ser105, Asp187, and His224 residues, and the substrates. The difference for the binding energy clearly points out that the enantiopreference of lipases toward organic compounds is capable of predicting by employing computational simulations with quantum chemical calculations.

View full abstract

Hydrolysis reaction is one of the major sources of degradants in the biodegradation test of Chemical Substance Control Law in Japan. We intended to add a new view in the biodegradation prediction using activation energy of hydrolysis reaction derived from the reaction analysis through quantum chemical calculation. In the process of above study, we developed a program, Hydrolysis Reaction Analysis System as an interface enabling easy calculations of activation energy by using MOPAC program. Substitution method was applied as an effective method to construct initial molecular structures for reactants, products and transition states of the ester hydrolysis. The developed system calculated activation energy of the acid hydrolysis reactions of esters with drastic reduction of computation time in comparison with those using other molecular modeling programs. If more precise calculation of the activation energies are required, it could be a solution to utilize the higher level of theory such as Density Functional Theory or ab initio MO calculation.

View full abstract

Molecular simulations using the conventional force fields have been successful for understanding the behavior of molecules in solvent. However, the classical approaches have some shortcomings due to the electrostatic interaction calculated by fixed atomic charges. In previous study, we developed the new charge equilibration approach including solvent effect by generalized Born (GB) formula (QEq-GB). This new method is available for simulations of biomolecules because the induced polarizations by both geometrical change of molecules and solvent can be included simultaneously. In this study, we introduced the two other ways of calculations for effective Born radius and investigated the dipole moments and charge distributions obtained by these methods. The results indicated that some modifications for parameter sets are necessary.

View full abstract

The conversion efficiency of dye-sensitized solar cell depends on the structure of the dye, and much effort is devoted to the development of dyes that give high Isc and Voc. The data of dye structure and conversion efficiency for 978 dyes, obtained in the primary screening process in Nihonkayaku, are used to extract characteristic substructures that give high Isc and Voc values. Mining method consists of the generation of linear fragment descriptors, analysis by the cascade model and decision tree and refinement of the motif substructure. The last step of the extraction process is done by a chemist's inspection on supporting structures. Three efficient substructures were found from the analysis of Isc values, and they covered 11 compounds out of 20 with the highest Isc values. All obtained substructures were characterized by the presence of a stable resonance structure with the intramolecular charge transfer.

View full abstract

In order to evaluate pi-conjugation effects on intramolecular resonance-assisted hydrogen bonds DFT calculations were carried out on 1-carbonyl and w-hydroxyl substituted polyacenes fused with 3-, 5-, or 7-membered pi-conjugated ring. The H-1 NMR chemical shifts and the Mulliken charge densities of the hydrogen-bonded protons, the electron densities at the bond critical points of intramolecular O...H hydrogen bonds, and the stabilization energies due to the hydrogen bonds were found to be linearly correlated well with the O...O hydrogen-bond distances. Whereas the values for the H-1 chemical shifts and the Mulliken charge densities of the hydrogen-bonded protons of the acenes with the fused 3- and 7-membered rings were larger than those of the corresponding acenes with the fused 5-membered rings in the region with smaller pi-conjugation bonds, the former became smaller than the latter in the region with larger pi-conjugation bonds. Similar changes of the values for the HOMA indices were obtained. These results suggested that the intramolecular resonance-assisted hydrogen bonds with (4N+2) pi-conjugated electrons were strengthened in smaller pi-conjugation bonds, but that those with 4N pi-conjugated electrons were strengthened in larger pi-conjugation bonds.

View full abstract

To elucidate the stability and the structure of a molecule, it is important to take account of the flexibility of the molecule. An equilibrium structure is a local minimum structure at the bottom of the potential energy surface; it differs from an average structure which is related to an observed dynamic structure in real systems. For different molecules which have similar structures, equilibrium structures may be similar, but average structures and the flexibility of molecules may differ. Here, octane and 1,2-diethoxyethane, either of which has a linear structure with 8 heavy atoms, are taken as the objective molecules. Equilibrium structures are calculated in the framework of ab initio MO method, and average structures and the flexibility of molecules are calculated using direct ab initio MD method. There are a lot of conformers of octane and 1,2-dithoxyethane in a small range of energy with low barrier heights. Based on the MD trajectories, we find that the average structures and the flexibility of both molecules are different, although the equilibrium structures are similar.

View full abstract

3-(Benzylthio)thieno[3,4-b]indolizine derivatives show significant spectral changes depending on the solvent. We have examined the possible factors, reaction and isomerization for the change. Geometry optimizations were performed at the B3LYP/6-31G(d) level. Based on the obtained geometry the electronic transitions were predicted by the TD-B3LYP calculation. Potential energy curves as for the dihedral angle change of the exocyclic sulfide linkage were examined. The barrier height from gauche to trans isomer was calculated to be about 5 kJ mol-1. TD calculation shows that the protonation on the carbonyl oxygen effectively changes the absorption spectrum.

View full abstract

Theoretical studies of van der Waals complexes provide valuable information about the nature of weak intermolecular forces. Last year, we reported the ab-initio study of the potential energy surfaces of C2H6/H2O and C2H5F/H2O Van der Waals dimmers. In C2H6/H2O, one stable structure was found, and, in C2H5F/H2O, two stable structures were found. In the present study, we have investigated the potential energy surface of C2H5F/H2O in more detail. The computed potential energy surface exhibits four stable structures. Two of them (referred to as ST1 and ST2, respectively) are those in which the H2O molecule exists in the C-C-F plane of C2H5F. The others (referred to as ST3 and ST4, respectively) are those in which the molecular plane of H2O molecule is perpendicular to the C-C-F plane of C2H5F. Both ST1 and ST3 have the H2O molecule in an equatorial location of the C-C axis, and both ST2 and ST4 have the H2O molecule in an axial region of the C-C axis. The details of the potential energy surfaces are discussed.

View full abstract

Intermolecular interactions play a significant role in the field of physics, chemistry and biology. The study of intermolecular interactions is a useful step toward gaining a basic understanding of the dynamic behaviour of molecules in various environments. In the present study, high-level ab initio calculations [MP2] using correlation consistent basis sets of Dunning [aug-cc-pVDZ, aug-cc-pVTZ] were carried out to study the potential energy surfaces of CH3OCH3/H2O and CH3OCF3/H2O Van der Waals dimers and to clarify the stable structures of these clusters. For the calculation of interaction energies, the basis set superposition errors (BSSE) were estimated by counterpoise method. One stable structure for CH3OCH3/H2O dimer has been found. For the optimized structure, the H2O molecule is outside the C-O-C plane of the CH3OCH3, and the bond angle O(CH3OCH3)---H-O(H2O) is approximately 180°. On the other hand, for the stable structure of CH3OCF3/H2O, the H2O molecule exists in the C-O-C plane of the CH3OCH3/H2O molecule. The O-H(H2O)--F(CH3OCH3) interaction is considered to be significant. The details of the potential energy surfaces are discussed.

View full abstract

We analyzed geometrical isotope effect (GIE) by substitutions of proton(H)/deuteron(D) in water-ammonia clusters as a model of interactions between surfaces of a protein and surrounding water molecules. Multi-component molecular orbital method with Hartree-Fock level of theory (MC_MO-HF) was applied to treat quantum effect of nuclei explicitly. Single s-type GTF was used for each proton and deuteron with 6-31G(d,p) electron basis set. Trimers composed of water and/or ammonia molecule works as an acceptor and a donor of hydrogen for each hydrogen bond were considered in this study, comparing with corresponding monomers and dimers. As for monomers, the bond lengths of X-H(D)(X=O,N) by MC_MO method were longer than those by conventional MO method because of quantum effect. The bond lengths of X-D were shorter than those of X-H due to anharmonicity of potential. Concerning dimers and trimers, the distances between O(N) showed the same order as increasing number of ammonia molecules. The bond lengths of X-H(D) derived from MC_MO method were longer than those derived from conventional MO method, while the bond lengths of H(D)...Y(Y=O,N) and the distances between O(N) derived from MC_MO method were shorter than those derived from conventional MO method. The bond lengths of X-D and D...Y were shorter and longer than those of X-H and H…Y, respectively.

View full abstract

The extension of the visualization system of atomic and molecular orbitals -MOOTIC- We are developing a 3D volume rendering system, Molecule Orbital Observation Tool with Iso-surface and Cloud (MOOTIC) in order to grasp the characteristics of the atomic and molecular orbitals. This time we add the facility to see the overlap of the atomic orbitals residing in the atoms arbitrarily allocated. We can see the changes of the overlaps as the position of these atoms changes, depicting the surface of the atomic orbital with rendering by iso-surface in which 90% of a single electron are included. Using this system, we would discuss the relations between the overlaps and the equilibrium nuclear distances of the dimers including transition metal atoms,for examples Mn-Mn,Cu-Cu and Mn-Mn. We also discuss the possibility to use the the equilibrium nuclear distances and resulting overlaps in order to see the characteristics of chemical bond.

View full abstract

A lot of chemical compounds have been used to maintain our comfortable life even though some of them are environmentally harmful. The Chemical Substances control Law of Japan requires prior evaluation of certain hazardous properties of new chemical substances intended for industrial use. It is, however, very difficult to experimentally confirm whether or not all the chemical compounds easily degrade under the environmental conditions. As theoretical method is one of the powerful tools to investigate reaction mechanism such as hydrolysis of oxiranes listed in the law, the method is a candidate to avoid experimental difficulties. In order to obtain basic data for acid catalyzed hydrolysis of oxirane derivative, the reaction mechanism was investigated by using density function theory (DFT) calculations. An attempt to forecast hydrolysis of oxirane derivatives was done by comparing the activation energies (Ea) from experiments with calculatied ones.

View full abstract

A Cycl[3.2.2]azine molecule ([2.2.3]cyclazine) has a rigid C2v plane structure with a sp2-hybridized carbon perimeter clipped by a central nitrogen atom, as an isoelectronic analogue of a [10]-annulene. Various types of cyclazines have been reported and attracted much attention from physicochemical, magnetic, biological as well as graph-theoretical viewpoints. Despite their interesting properties, no systematic quantum chemical analysis have been reported to characterize the UV/Vis intense electronic peaks and the associated sidebands. In this study, semi-quantitative (TDDFT) and quantitative (MS-CASPT2) analysis are presented for a cycl[3.2.2]azine, its benzo-annulated and the maleimide-linked derivatives. Furthermore, the detailed spectral structures were analyzed on the basis of the orbital vibronic coupling interaction scheme.

View full abstract

We have been trying to use activation energies (Ea) as an index for degradation of esters in the natural conditions. However, it was confirmed that Ea is not the suitable index for all the esters cited in the Chemical Substances Control Law of Japan. On the other hand, activated free energies (ΔG‡) are directly connected with observed reaction rate constants. Instead of Ea, therefore, we intended to use ΔG‡ calculated from molecular orbital (MO) calculations including vibration frequency calculations at the MP2/6-31+G(d,p) level of theory and/or those from the MO/MC simulations (Monte Carlo simulations using MO calculations). In the MO/MC method, we used a droplet model that includes a solute molecule in its center and arranged 32 solvent molecules around the solute with in 6.5 A from the center. The calculated ΔG‡ values for methyl acetate is well consistent with that from the experiments. In addition, we calculated activation free energies (ΔG‡ obs) that come from actual measurements of the acid-catalyzed hydrolysis of esters to compare these values with calculated ones.

View full abstract

Catechins are antioxidants. catechins display anticancer activity, too. Recently, Kuzuhara et al. reported that epigallocatechin gallate (EGCG) binds both DNA and RNA molecules. Their experiments revealed that the branching structure of catechin, consisting of galloyl and catechol groups, is necessary for the interaction between DNA and catechin. In this work, we study the interaction between catechin and DNA (RNA) theoretically. Ab initio molecular orbital calculations are performed for EGCG with bases (9-methyladenine (9-MA), 1-methylcytosine (1-MC), 9-methylguanine (9-MG), 1-methylthymine (1-MT) and 1-methyluracil (1-MU)). We obtained an optimized structure for the complex of EGCG and 9-MG. We also obtained two kinds of optimized structures for the complex of EGCG and 1-MT. It was shown that the main interaction between EGCG and a base is hydrogen bonds. The stabilization energy (SE) for the complex of EGCG and 9-MG is -34.33 kcal/mol and the SE for the complex of EGCG and 1-MT is -29.95 kcal/mol. These are compared with the SEs for the Watson-Crick base-pairs. SE for A-T pair is -23.13 kcal/mol and SE for G-C pair is -40.00 kcal/mol.

View full abstract

The capillary gas chromatograph column retention time under isothermal conditions was observed for 46 pentachlorobiphenyls (PeCBs). Relative changes of the partition enthalpy (Δ(ΔH)) and entropy (Δ(ΔS)) were calculated on the basis of van' Hoff relation. A marked linear relationship between them, so called classical enthalpy & entropy compensation, was observed. A clear decrease of Δ(ΔH) was found with the decrease in the number of chlorine atoms substituted at 2, 2', 6, and 6' positions. The decrease indicated the enhancement of the interaction between the liquid phase in the column and the PeCB molecule, and seemed to be related to the planarity of the molecule. Furthermore, the difference of Δ(ΔH) from the linear relationship (ΔH-LFER) was investigated in terms of the number of a pair of proximitive hydrogen atoms (Npx-H). Negative difference was observed when Npx-H was 0 or 1, and positive difference was observed when Npx-H was 3 or 4. In addition, almost linear dependence of ΔH-LFER on Randic''s 2χ, known as a typical connectivity descriptor, was also revealed. To characterize the ΔH-LFER, which is discrepancy from the linear free energy relationship, quantitative evaluation of the effect induced by the atoms substituted at proximitive positions will be an important part.

View full abstract

Concerns about the labeling of agricultural or fisheries products at production centers is now increasing. After the outbreak of bovine spongiform encephalopathy (BSE) occurred, problems relating to clam or wakame and other fisheries products have are now attracteding more attention than before. We developed a scientific method for discriminating the production countries of wakame seaweeds using RandomForest. RandomForest is so-called a member of CART families. In this paper, we call cart families (CART, MARS, RandomForest and MART) as forest learning method (FLM) because we are able to bring to mind the concrete image of those methods. We found that RandomfForests could be a potent classification method for visible and near infrared spectra. Much more, we suggest a preliminary interpretation for the basis of classificationof classification conditions by using CART.

View full abstract

Artificial neural networks (ANN) have widely been used in predicting physical properties of chemical materials. In this work, an ANN analysis is applied to determining dielectric constants of various pure solvents (i.e. water, chloroform, acetonitrile, methanol, formamide, etc) and aqueous mixed solvents (dioxane-water etc). A solvent dielectric constant is a physical property which is quite difficult to measure precisely. We have already reported that the circular dichroism spectrum of an optically active cobalt(III) complex is effected by solvents used. By using peak positions (nm) and CD intensities of the cobalt(III) complex as a sensor molecule in various solvents, we tried to predict solvent dielectric constants. Multilayer neural networks were chosen for pattern-recognition analysis. The architecture consists of input, output, and hidden layers, which are where data are processed through a back propagation training algorithm. Of the 14 spectra measured for pure solvents, 13 of these were used to train the data set (peak positions (nm) and CD intensities) obtained from each spectrum. When the remaining one unknown is processed, an absolute error of a few % was obtained for the output solvent dielectric constants.

View full abstract

A method called Regression Discrimination Analysis (RDA) has been developed for the discrimination of ordered categorical data. The method (RDA method) has the advantages of simultaneously considering any numbers of classes and faster computation than other methods. However, the RDA method is not utilized because the RDA method is inferior in recognition and prediction. The RDA method has two problems. First, although the RDA method needs parameters called class values, the class values are not specified. For this reason, the most adequate class values have to be determined for each time. Secondly, the borders of the RDA method are also not specified. The change of the borders has great influence on the result of discrimination. In this study, we suggest the optimized class values which are determined by searching for the normality in the data belonging each class value. We have refined the RDA method to some extent in this study.

View full abstract

Kohonen-type Self-organizing map (SOM) is widely used in clustering. When using SOM, you can have a trouble in choosing proper parameters. Generative topographic mapping (GTM) proposed by Bishop et al. seems to be superior to SOM, because in GTM there is a well-defined objective function given by the log likelihood. And the objective function convergence to a (local) maximum is guaranteed by the use of EM algorithm. In this paper, we made a comparison between these two mappings by the both artificial and real data, and also evaluated the mapping performance of both GTM and SOM by introducing the root mean square of midpoint (RMSM). RMSM is calculated as RMS of mapping error of all midpoints instead of the original data points. RMSM represents how the map complement between data points. Comparing RMSM, you can know which technique maps more smoothly onto the data space. We applied this barometer to the wine data. It indicated that GTM produces the smoother map than SOM, and the RMS of GTM is approximate to that of SOM. Thus GTM can be superior to SOM as mapping method.

View full abstract

Regression analysis using chemometric methods is effective to improve efficiency of research and development. When regression models to predict objective variables by design variables are built, it becomes possible to optimize the design variables by use of the models. If the models are interpretable, there is some chance of discovering new knowledge. In this work, chemometric methods are applied to automotive catalyst. We have analyzed data about TWC (three-way catalyst) and DPF (diesel particulate filter) as case studies. The relation between compositions of TWC supports and corresponding values of catalytic activities were modeled by BPNN (back-propagation neural networks). We optimized the compositions of TWC supports by using the model and found new compositions that are expected to show high catalytic activities. Secondly, We built a PLS (partial least squares) model to predict maximum temperature during regeneration of DPF for the purpose of keeping DPF from cracking by thermal shock. It seems to enable robust design of DPF and stable control of regeneration.

View full abstract

A lot of researches of Mn salen catalysts which are capable of epoxidizing various olefins have been reported because they are able to be synthesized from one ethylenediamine and two salicylaldehyde through dehydration reaction, and many derivatives with characteristically steric structures are obtained easily. For the synthesis of catalysts with high enantioselectivity, it is necessary to understand how the catalyst reacts with olefins especially in their geometrical relation and which substituent is effective for successful reaction. Though theoretical calculation such as quantum chemistry would be a useful tool to investigate these mechanisms and search their molecular structure or several physicochemical properties, it is difficult to use this theoretical approach for quantitative prediction of enantioselective reactions because the energy difference between intermediates that lead to enantiomeric products is small. In this study, QSAR approach, mainly CoMFA method, was applied to estimate enantiomer excess in asymmetric epoxidation of olefins using Mn salen complexes as catalysts.

View full abstract

In QSAR studies, it is known that a model using variables irrelevant to object variables often have a low predictive ability. So in order to build a model that gives high predictive ability, it is important to select minimal number of the variables necessary for prediction. Many variable selection methods have been developed, and we propose the new method using rough set theory (RST). RST is the theory that is used to classify inaccurate or incomplete data sets. By applying this theory to variable selection, we can find minimal subsets of variables which give the same partition as the whole set of variables (reduction set). In this study, this method is applied to QSAR analyses of inhibitors of DHFR and HIV-1. We built the PLS models both using reduction sets and using variables randomly selected. PLS models made by reduction sets give predictive power not less than predictive power which the models made by randomly selected variables give. Furthermore, it became cleared that chemically meaningful variables are tend to be selected by RST, and therefore, shown that RST is superior to the random selection. After this, we will continue to examine the advantage of RST by comparing with other methods.

View full abstract

A parallel parameter optimization for Support Vector Regression (SVR) is proposed. A central of this technique is the use of population-based search, called Particle Swarm Optimization (PSO), for the simultaneous optimization of multiple parameters in SVR modeling. PSO is a biologically inspired search algorithm motivated by a social analogy. It is an iterative method based on the search behavior of a swarm of particles "flying" through a multidimensional search space. The particles have two biological properties as follows; they can remember the position (i.e. candidate for the solution) they have ever searched and can also share the information related to the solution to be solved and, consequently, these biological properties make it possible to reach global solution of the problem. In the present study, optimization of SVR parameter such as loss function and parameter of Gaussian kernel function is demonstrated with a QSAR data set. An optimization of robust kernel based on Huber's M-estimator is also shown.

View full abstract

Bioactive peptides, e.g. neuropeptides and peptide hormones, often have post-translational modifications like C-terminal amidation or N-terminal pyro-glutamination. A new peptide ghrelin, which found in 1999, had been found that its 3rd amino acid residue serine is octanoylated. To discuss these peptides from the viewpoint of physicochemical property and activity relationship, we have no hydrophobic parameters covering these modified amino acids. We expanded ClogP calculation concept to medium-sized peptides with parameters divided into each residue contribution. These parameters can be summed up along amino acid sequences, and provide exactly the same values as those of ClogP without drawing any SMILES structural expressions. We showed two examples, thyroliberin and ghrelin calculations in this paper. This method provides a consistent way of hydrophobisity prediction from small organic compounds to large proteins. This idea named Biopolymer ClogP (BP-ClogP) can be expanded to other bio-molecules like sugar chains, glycoproteins, lipids, lipoproteins, DNA and RNA molecules.

View full abstract

The concept of drug-likeness is important on designing chemical libraries and on virtual screening. The concept is rather subjective and in the early applications, the drug-likeness of compound is determined by medicinal chemists. Although the discriminating is mainly carried out by using machine-learning software nowadays, the ideal training set used is not clear. In order to construct a good example of drug-likeness discrimination, we plan to collect discriminated data done by multiple and diverse medicinal chemists. As the in-house and/or commercially licensed chemical data and commercial software are used in the most of past studies, the discriminators of the data are very limited. In this study, a web-based drug-likeness evaluating program is developed only uses open source software such as MyFaces, MySQL and The Chemistry Development Kit. The compounds to distinguish are selected from the 250,251 data set collected by NCI. Therefore, the program can be used by any chemists without license limitations. The program can try via our web site. The program itself will be available from our web site soon.

View full abstract