Thermodynamics in the condensed phase is a challenging target of the present quantum chemistry. Because the widely-spread quantum chemical programs such as Gaussian and GAMESS adopt the ideal gas model (IGM) even for the liquid systems. We have recently proposed a novel approach, named harmonic salvation model (HSM), to properly treat the condensed phase. In the presentation, I will explain the theoretical aspects and introduce several applications of HSM.

View full abstract

Data assimilation is an interdisciplinary science to connect numerical simulations and real-world data and plays a central role in numerical weather prediction. Both simulations and sensor technologies keep advancing, and data assimilation deals with these large-volume and high-velocity "Big Data". Real-time processing of Big Data from high-precision simulations representing individual clouds and the new phased array weather radar data that capture rapidly-changing dynamic weather every 30 seconds is a real challenge. The revolutionary "Big Data Assimilation" technology aims to predict precisely where, when, and how intense rainstorms will occur.

View full abstract

Crystal structure prediction provides useful theoretical knowledge for designing and synthesizing new materials in pharmaceutical therapeutics and industrial electronics. Therefore, we have developed a high-speed and high-accurate prediction method for organic molecular crystal structures. In this work, in order to consider effectiveness and problems in the prediction method, we perform the theoretical predictions of crystal structures of fourteen target molecules using our method. Furthermore, we have developed a master-worker type parallel method for the crystal structure prediction.

View full abstract

We have been developing a data base named QMRDB which shows the data in Web browser in conjunction with the PostgreSQL program. It is possible to perform not only the data search using characters such as molecular names, keywords but structure matchings using the implemented function of the OpenBabel program. The Jsmol program displays the structure in a web browser together with molecular names, keywords, energies and so on. QMRDB was applied to develop Transition State Data Base (TSDB) in the present study. As we adopted a specific description suitable for describing chemical reactions for TSDB, their details and usages are reported.

View full abstract

One of the most fundamental questions in chemistry is how does structurally modifying a molecule with substituents affect thermodynamic properties of a condensed molecular system. Itoh et al. reported a reduced viscosity in a fluorinated derivative of the 1-butyl-3-methylimidazolium (bmim) salts. At first, we suspected that the reduced viscosity was caused by modification of the molecular motion induced by the fluorine atoms, because the rotational potential calculated for the fluorinated bmim cation was different from that for the bmim cation. Unfortunately, the difference of the rotational potential could not be ascribed to the reduced viscosity. This implies that the effect of the substituents on the thermodynamic properties should be studied with the same approach as many body problems. To answer the question, we defined the molecular conformations by Potts spin variables as a general coordinate. The mean field Hamiltonian was formulated as a quadratic form of the Potts spins and applied to 1-methyl-3-propylimidazolium chloride (pmimCl) derivatives. Now we report the temperature dependence of the order parameters and the relationship with the conformational energies in the pmimCl derivatives.

View full abstract

In the molecular animation production for visualizing in vivo　molecular reacitons, a static structure of protein model is sometimes　inadequate to consider dynamic aspect of the molecules. We will　discuss about our new method to describe the dynamic motion of　proteins including the thirmal fluctuation based on the normal mode analysis of proteins.

View full abstract

It is very difficult to visualize wave function because it is four-dimensional data consisting of the value in the three dimensional physical space. We visualize the information by projecting the wave function onto two-dimensional subspace. In this study, we improve effective retrieval of molecular quantum information using the method with a focus on the subspace projection and real-time processing.

View full abstract

There has been almost no book and textbook of organic chemistry that stress the importance of the concept of cross-conjugation in unsaturated conjugated systems. The author has been analyzing this problem by using the graph-theoretical molecular orbital method, and clarified that this concept is playing an important role for verifying and understanding the essence of the electronic theory of organic chemistry. Further, he could also show the limit of this theory for actual application. The present author’s theory can be extended to another important problem of aromaticity and anti-aromaticity. Although this theory is based on the conventional HMO method, the essence of the present result will not be changed even by elaborating the level of the theory.

View full abstract

Solvent viscosity dependence of reaction rate is formulated by Kramers and Grote-Hynes. We succesfully reproduced Kramers turnover of reaction rate of azobenzene Z/E-isomerization using metadynamics methods.

View full abstract

Recently, attention is being given increasingly to photomechanical materials for which the shape change of photochromic molecules can be extracted as motion when irradiated. Using molecular dynamics and the Monte Carlo method, we investigated physical and chemical aspects of the interaction between an oligomer and a photon. Additionally, we observed and discussed the motion of oligomers.

View full abstract

Amyloid-β peptides (Aβs) play a key role in pathogenesis of Alzheimer's disease. Since the aggregation of Aβs is involved in the pathogenesis, the conformations of Aβ aggregates have been investigated by many experimental and computational studies. In the present study, we performed replica exchange molecular dynamics (REMD) calculations to obtain various conformations of full-length Aβ(1--42) dimer in water and determined the most stable conformation of the solvated Aβ dimer by ab initio fragment molecular orbital (FMO) calculations. In the most stable conformation elucidated by the present study, β-hairpin and zipper-like β-sheet structure are formed, being consistent with the previous studies. We furthermore investigated the specific interactions between Aβ monomers by FMO calculations, in order to elucidate which residues contribute mainly to the stability of the Aβ dimer. As the result, it is elucidated that Lys16 and negatively charged residues of N-terminal contribute mainly to the stability of the most stable conformation of solvated Aβ dimer. In addition, the present study molecular simulations with solvating water molecules considered explicitly elucidated that some solvating water molecules contribute significantly to the stability of the solvated Aβ dimer.

View full abstract

In the glycoscience, there are various styles for representing glycan data, and that have been an obstacle to link life science data using the Semantic Web technologies. In order to solve this problem, we have developed Web3 Unique Representation of Carbohydrate Structures (WURCS) as a new unique linear notation for representing carbohydrates for the Semantic Web. WURCS has the following features: linear notation so that it can be used as a URI, and a unique notation such that any published glycan structure including ambiguity can be represented distinctively. WURCS can be used to search for the same carbohydrate moiety in databases implementing this format. Furthermore, we developed a tool that generates WURCS using a structure-drawing editor. We also organized a WURCS Working Group for the promotion of WURCS, which currently involves glycoscientists from around the world.

View full abstract

Tetrahedral, allene, ethylene, oxirane, and square planar skeletons are discussed from a viewpoint of the group hierarchy, which consists of point groups, RS-stereoisomeric groups, stereoisomeric groups, and isoskeletal groups. To treat these groups comprehensively, the isoskeletal group is equalized to a reflective group based on the symmetric group of degree 4. The proligand method is extended to accomplish combinatorial enumeration under these groups. The influence of these groups on stereochemical nomenclature is discussed.

View full abstract

We extend our previously developed enumeration methods in breadth first search order, and propose an algorithm for enumerating all non-redundant chemical compounds containing naphthalene rings as cyclic structures. We deal with a naphthalene ring as two special atoms with a special bond, and enumerate tree structures instead of general graphs. For evaluation of the performance, we compared the computational time with that of MOLGEN, a well-known enumeration tool. The results show that our algorithm is much faster than MOLGEN.

View full abstract

Molecular biological data has rapidly increased with the recent progress of the Omics fields, e.g., genomics, transcriptomics, proteomics and metabolomics that necessitates the development of databases and methods for efficient storage, retrieval, integration and analysis of massive data. In the present study we introduce KNApSAcK Family DB to comprehensively understand species-specis interactions integrated by traditional and modern knowledge and Omics fields.

View full abstract

It is reported that a portal web site for information on clinical research (trials) of National Institute of Public Health (NIPH) which provides clinical trial information in Japan is hard to use, especially difficult to find the information they want, for nonprofessional users including patients. To improve the retrieval capability of the site, we first tried to find problems on the site. For comparison, we also investigated the retrieval capability of ClinicalTrials.gov. In the case of Japan synonyms of the query cannot be found. Although they can almost be found in ClinicalTrials.gov, it is found that the coverage isn’t complete. Moreover, both of them are not able to retrieve hypernyms and hyponyms of the query. We considered that the retrieval system including synonymy, hepernymy, and hyponymy is necessary for nonprofessional users to get all the information they need. Using dictionary whose words are extracted from Thesaurus of Medical and Health related Terms, we then developed a retrieval system including synonymy and hyponymy. As a result, it is confirmed that the system can retrieve a wide range of useful information for the public and patients.

View full abstract

Simulation of the dynamic behavior of the metabolites requires initial amounts of metabolites and all the reaction rate constants to be known. However, it is difficult to obtain actual amount of the metabolites in the experiment from a mass spectrometry, and reaction rate constants are not always known. In the present study, we develop a tool to estimate those parameters that can reproduce the metabolome data obtained from the experiment using a distributed genetic algorithm. We used the parameters to a stochastic simulation, and then reproduced the experimentally observed dynamics of amino acid biosynthesis in Arabidopsis thaliana. Parameters estimated with a distributed genetic algorithm can reproduce more appropriate behavior of the metabolism than that with a genetic algorithm. This method can provide users an estimation of the dynamic behavior of the metabolites whose data are unavailable from the experiment, as well.

View full abstract

We discuss applicability domains (ADs) based on ensemble learning in classification and regression analyses. In regression analysis, the AD can be appropriately set, although attention needs to be paid to the bias of predicted values. However, because the AD set in classification analysis is too wide, we propose an AD based on ensemble learning and data density. First, we set a threshold for data density, below which the prediction result of new data is not reliable. Then, only for new data with a data density higher than the threshold, we consider the reliability of the prediction result based on ensemble learning. By analyzing data from numerical simulations, we demonstrate that the ADs based on ensemble learning are too wide. Then, by using quantitative structure-property relationship data and quantitative structure-activity relationship data, we validate our discussion on ADs in classification and regression analyses and confirm that appropriate ADs can be set using the proposed method.

View full abstract

A chemoinformatics system is developed for molecular search. The system consists of molecular database storing computational results of electric-structure calculations, a knowledge search tool using text-mining technique, molecular similarity analysis tools, and data-mining tool using statistical analytic methods. Based on the calculated energy spectra of molecules, molecular similarity is quantitatively evaluated. In our chemoinformatic system, important electronic factors are automatically suggested through automatic (machine-learning-type) comparisons of experimental data with the accumulated computational data. Literature database is combined with the present database by using a text-mining tool in which important keywords of molecules are extracted from electronic files of scientific papers. The present system is applied to discover tyrosinase inhibitors.

View full abstract

Ethanolamines are known to absorb the carbon dioxide and attracted much attention in these years. Ethanolamines have several conformers within their backbone structures. We have performed the geometry optimizations of the methanol amine and four water molecules and found many structures that are suitable for the proton transfers.

View full abstract

Because the hydrophobic effects play important roles in the formations of drug-target complexes, stable and bulky hydrophobic groups are useful for drug design trials. Carborane, which is one of the boron cluster, is expected to be useful for drug design trials as the novel hydrophobic moiety. On the other hand, force field parameters for boron are not sufficiently developed. Because the force field parameters are indispensable to classical molecular mechanics calculations, in silico drug design trials using classical mechanical methods were occasionally inapplicable to carborane derivatives. In this study, force field parameters of carborane moiety were determined using quantum chemical calculations. Because the bond lengths and angles cannot change independently in the carborane structure, multiple regression analyses were used. The atoms in carborane were slightly moved, and the changes of energies were calculated by quantum chemical calculations. The multiple regression analyses using structural features, e.g. angles and bond lengths, as independent variables and quantum chemically calculated enegies of the structures as dependent variables were carried out to determine the force field parameters.

View full abstract

We have revealed the electronic structure of the ground and excited states of the DyF molecule below 3.0 eV by applying the 4-component KRCI method. About 100 million Slater determinants are generated by single and double excitations from the reference space composed of the Dy (4f), (6s), and (6p) spinors. We have performed the CI calculations by the KRCI program parallized for the multi-core processor. The dominant configuration of the ground state (X(1)7.5) is (4f)9(6s)2 and those of the 4 low-lying excited states are (4f)10(6s)1. McCarthy et al. assigned the dominant configurations of the [19.3] and [20.3] states (at 2.39 eV and 2.52 eV) as (4f)9(6s)1(6p)1 from the spectroscopic study and the computation based on the ligand-field theory. However, we assinged them as (4f)10(6p)1 since their large Re's (3.88 au) are consistent with experiment.

View full abstract

It is well known that the amino acid sequence of a protein is closely related to its structure and function. In the present work, the authors have proposed a representation of protein sequence features based on combination of a target residue and its symmetric neighborhood residues, and constructed the comprehensive palindrome fragment database based on NCBI RefSeq (HUMAN) which contained 35,665 entries (the total number of residues: 19,864,982). All palindrome fragments for each size were hierarchically stored with the information of the target residue, the total number of occurrences, the number of entries, and the number of constituted amino acid type. For size-5 fragment such as “PPGPP”, 6,213 / 8,000 patterns (total 157,481 fragments) were obtained from the database. Sequence feature analysis was performed using the PROSITE motif patterns and a frequency of each fragment in the database. As a result, we successfully found the characteristic palindrome fragment which is closely related to the common protein family.

View full abstract

Divalent ions play important roles in biological systems, such as in various proteins. Studying hydration structures and hydration energies of metal ions is important to elucidate how metal ions function in proteins. In this study, we focus on divalent ions such as Ca²⁺ and Mg²⁺ .Cluster hydration structures and two-body interaction energies are calculated using ab initio MO method. Hydration structures in droplet model are calculated using QM/MM method. We reveal that the hydration structures of metal ions in cluster models are not always same as those in droplet model.

View full abstract

Emergence of drug resistant viruses caused by treatments of viral infections is inevitable and the rapid evaluation of drug resistance is a significant subject as a countermeasure. Previously, we developed a novel method for evaluating drug resistance abilities of pathogenic viruses by a docking study, using known crystal structures and the mutation information. In this study, we applied this method to evaluate the abilities of atazanavir as well as oseltamivir in order to extend the scope of this method. First, we tried to evaluate the resistance abilities of atazanavir against 30 kinds of HIV-1 proteases, including a wild type and 29 variants. As a result, it was shown that our method was able to correctly evaluate the highly resistant protease. Secondly, we also applied this method to subtypes H1N1, H3N2, and H5N1 as well as type B neuraminidase, which include the wild type, low-level and high-level resistant variants. Then it was shown that our method was also able to correctly evaluate the highly resistant variants of subtype H1N1 and type B neuraminidase.

View full abstract

H₃⁺, H₃ and H₃^- are the simplest homo nucleus three atomic molecules and examples of three center two/three/four electron bond respectively. The structure of H₃⁺ takes the equilateral triangle that three H atomic nuclei become equivalent, because two electrons of H₃⁺ occupy the first node-less σorbital. In the case of H₃ and H₃^-, an electron (or two electrons) occupies the second σorbital which has a node through an H atom and a center of H-H bond of the equilateral triangle of H₃⁺. Thus, the structures of H₃ and H₃^- become linear. In this note, we show the angular dependence of molecular structure, orbital energy, total energy and Mulliken charge of H₃⁺, H₃ and H₃^- using HF/6-311++G**.

View full abstract

In drug design study, it is often observed that a very small structural modification such as introducing a methyl group or a chlorine atom to the known inhibitor causes large enhancement of the inhibitory activity. However, it is difficult to predict such an activity change from structural information even if the structure-based drug design method is utilized. Therefore, selection of such compounds guided by calculated binding energy from candidate compounds is one of the efficient ways. In this study, we employed the thermodynamic integration method for the prediction of binding free energy change between two compounds to explore more potent CK2 inhibitors by introducing a chlorine atom to the known inhibitor. CK2 is a kind of serine/threonine protein kinases and is a target protein for the treatment of cancer and nephritis. Since a part of compounds were predicted more potent than the base compound, we synthesized and evaluated their inhibitory activities, and their structure-activity relationship was analyzed.

View full abstract

Ru(II)-Pheox complex was found to be efficient catalyst for intramolecular cyclopropanation to afford cyclopropane-fused γ-lactones in high yield with excellent enantioselectivity (up to 99% ee). In this work, the mechanism of Ru(II)-Pheox catalyzed intramolecular cyclopropanation has been studied computationally.

View full abstract

Generating pharmacophore models and screening compounds with them play an important role in lead discovery and optimization. However, without knowledge about its algorithm, especially when we generate models based on ligand superposition, it should be difficult to obtain a good predictive models. Here, I outlined the algorithm of pharmacophore modeling employed in BIOVIA Discovery Studio and pointed out suggested tricks and traps upon selecting compounds for model generation.

View full abstract

Recently, metal organic framework (MOF) has attracted much attention as new porous adsorbent materials. Thousands of MOF structures can be fully crystalized from the various combinations of organic and inorganic subunits. In the present study, we investigated the stable structures and their electronic properties of the model systems of MOF, using the first principles molecular simulations based on density functional theory (DFT). We employed the model structure of the smallest unit for MIL-101(Al) and MIL-101-NH_2 (Al) MOFs. In addition, the interactions between Al ions and the subunits of MOF were investigated, using ab initio fragment MO (FMO) calculations. We investigated the interaction energies between Al and BDC/BDC- NH_2 by use of FMO method. The interaction energy between Al and BDC- NH_2 is significantly larger than that between Al and BDC. Therefore, it is elucidated that the NH_2 introduction to BDC enhances the interactions between Al and BDCs, resulting in the more stable framework for MIL-101- NH_2 (Al). We furthermore proposed new BDC derivatives and investigated the stability of the MOF systems including them. The results will be presented at the conference.

View full abstract

Lactose repressor protein (LacR) controls the transcriptional mechanism of gene information from DNA to mRNA in a ligand-dependent manner. Although the ligand-binding to LacR was found to change the mechanism drastically, the effect of ligand-binding on the conformation of LacR+DNA complex has not been clarified at atomic and electronic levels. We here investigated the change in conformation of LacR-dimer+DNA complex induced by the ligand-binding, using molecular dynamics simulations and ab initio fragment molecular orbital calculation. The results elucidate that the binding of an inducer to LacR significantly changes the LacR structure to cause strong interactions between LacR monomers, resulting in weakening the interactions between LacR-dimer and DNA.

View full abstract

In recent years, DMET(Density Matrix Embedding Theory), a kind of simplified version of DMFT(Dynamical Mean Field Theory), has been proposed. Applying this DMET to the Hubbard model on the honeycomb lattice, we tried to understand the electronic state of graphene.

View full abstract

Butyrylcholinesterase (BChE) exists mainly at hippocampus and hydrolyses neurotransmitter acetylcholine. Many types of ligands have been produced to inhibit the activity of BChE. We here investigate the specific interactions between BChE and several types of ligands Kx, using ab initio fragment molecular orbital (FMO) calculations. The FMO calculations highlight the amino acid residues of BChE important for the binding between BChE and Kx. Based on the results, we propose some novel ligands and elucidate that one of the ligands can bind more strongly to BChE than Kx.

View full abstract

Recently, many incidents and accidents are caused by the use of dangerous drugs. As a countermeasure, 800 kinds of synthetic cannabinoids and 500 kinds of cathinones are comprehensively specified and regulated. There are many things of not synthesized in a lot of standards of these drugs. Therefore, it is difficult to identify by means of comparing these drugs and standards. We thought a method of mass spectral prediction about cannabinoids by use the density functional theory of quantum chemistry. In this way, we compared the mass spectrum of prediction by using quantum chemistry and those standards were measured by LC-MS/MS with ESI (Electrospray ionization). We will report of poster about these details.

View full abstract

Recent biochemical experiments have revealed that many proteases play important roles in cancer invasion and metastasis. Among these proteases, the binding of urokinase-type plasminogen activator (uPA) and uPA receptor (uPAR) existing on the surface of a cancer cell is a trigger for cancer invasion. Therefore, the blocking of the binding is expected to inhibit cancer invasion. In the present study, we created a various types of peptides for blocking the ligand-binding pocket of uPAR. The binding energies between uPAR and the peptides were evaluated by ab initio//fragment molecular orbital calculations, and the peptide with the largest binding energy was proposed as a potent inhibitor to cancer invasion.

View full abstract

To elucidate the dependence of geometric structure and thermal stability of liposomes on their component phospholipid molecules, we conduct theoretical simulations of the eight types of liposomes in water constructed from different phospholipids. By the use of molecular mechanics simulations based on coarse grained (CG) model, stable structures of the solvated liposomes are obtained. In addition, their thermal stability is investigated by the molecular dynamics (MD) simulations based on the CG model at 310 and 360 K. The results of 100 ns CG-MD simulations indicate that liposomes with the same number of lipids on both the inner and the outer layers keep their spherical form even at 360 K. The present results will be useful for developing novel liposomes with enhanced stability. For more information about the results of the comparison and structural changes with respect to the temperature change of the liposome, I will be announced at a research poster on the day.

View full abstract

In order to automate the research and development process for chemical vapor deposition (CVD) processes, we developed and proposed a concept of planning experimental design automatically. The concept was applied to identifying the reaction mechanism (reaction model) that indicates the reaction paths both in a gas-phase and on a surface from the reactant (source gas) to the product (film). In addition, we demonstrated the feasibility and validity of the concept which we proposed using a synthetic CVD process.

View full abstract

Recently, the influence of radiation on human body has been recognized as a serious problem. In particular, highly-reactive radicals produced by the radiation react with DNA, resulting in a great damage on its structure and electronic properties. It is thus important to investigate the reaction mechanism of radicalsto DNA for elucidating the initial damage in DNA induced by the radiation. In the present study, we search for the stable structures and transition states of the reaction for the complexes of G-C/A-T base-pair withradicals in vacuum and in water, by the density functional theory calculations.

View full abstract

Developing database systems connecting diverged species based on omics is the most important theme in big data biology. To attain this purpose, we have developed KNApSAcK Family Databases, which are utilized in a number of researches in metabolomics. In the present study, we have developed a network-based approach to analyze relationships between 3D structure and biological activity and applied to a data set consisting of 2,072 secondary metabolites and 140 biological activities reported in KNApSAcK Metabolite Activity DB. As a whole, we systematically analyzed the relationship between 3D-chemical structures of metabolites and biological activities.

View full abstract

We tried auto-classification of molecular structure for a simple example problem. As the input data, we used a part of atomic distance matrix extracted from the results of Gaussian 09 geometry optimization job. As the result, we could get reasonable classification rules automatically.

View full abstract

Spatial operations of a molecular model using a keyboard or a 2D input device " mouse " are complicated because it is necessary to switch their device modes. To improve this problem, a molecular model is displayed like to be present in the real space using AR technology, the displayed molecular model has been able to be spatial operations using " hands " . In this study, by using the " Leap Motion " which is a kind of the motion controller devices, we have implemented the distinctive spatial operations into AR Molecular Viewer that has been developed in our laboratory. As a result, by operating with the real " hand " motions, it has become possible to perform the rotating and scaling the molecular model on the display. In the future, we will continue to improve so as to approach the feeling that we are operating with a three-dimensional molecular model in the real space more.

View full abstract

Molecular polarizability is one of the optical resonance property. It is very important to compute frequency dependent (dynamic) polarizability of molecule at quantitatively in a resonance region. In this study, molecular polarizability is calculated using quantum chemical theory. For wide range of frequency, we use time-independent theory to compute the property. We introduce a damping factor in the theory in order to avoid divergence of polarizability at resonance region. In quantum chemistry, Linear Combination of Atomic Orbitals (LCAO) method has been widely used for its efficiency. But the accuracy of the calculation using LCAO basis is not mathematically controlled. We cope with the problem using Multi Resolution Multi Wavelet Basis set which provides the calculated property with guaranteed precision. In this study, we developed a program system which quantitatively evaluate the dynamic polarizability of molecule for wide range of frequency including resonance region using MRMW basis sets and the results are compared with the values of obtained using LCAO basis sets.

View full abstract

Everyone can obtain information of various molecules about electronic structure calculation by means of development of computational chemistry. However, searching with focused on necessary data is need for using efficiently those data. In this study, we tried to build the database, which can accumulate and arrange of calculation results. Moreover, we tried to equip the systems assisting for molecular design. Following systems were implemented as basic functions of database: (1) database, (2) search system, (3) assisting molecular design system, (4) electronic structure analysis system, (5) calculation result collection system and (6) performing molecular result system. Following systems were implemented for applying information accumulated by above systems: (1) molecular classifying system, (2) similar functional molecule search system and (3) text information search system.

View full abstract

An issue in development of functional molecules is to search a molecule which shows expected properties. A concept of molecular similarity is one of the strategies for searching molecules of interest. Because of its importance, various similarities have been suggested so far. In this study, we propose a method for quantitative evaluating molecular similarity on the basis of electronic structure calculations. Herein we suggest 9 types of descriptors for molecular similarity. They are calculated using electronic-structure calculations in information of electronic ground state, electronic excited state, and the vibrational modes of the electronic ground state are provided. We have applied our method to classify a set of organic molecules including 5 carotenoids and 11 hydrocarbons. It has been successful to distinguish carotenoids and other molecules without referring to their chemical formula.

View full abstract

A database is developed for storing chemical knowledge including chemical properties and application fields of a particular compound. It consists of knowledge table, between a compound and its knowledge information, derived from electronic literature on Internet. This data table is made in two steps: the first step is extraction of important keywords and compound names from a particular literature data. It has been confirmed that the selected keywords appear in the abstract section of the selected reference. The second step is to make a data table in which keywords for a compound are determined as a logical conjunction among the keyword groups included in the related references.

View full abstract

A new method of defining molecular shape is proposed. In this method, electronic structure calculation is applied to define the outer surface of a molecule given as an isovalue surface of the potential energy surface for interaction between a target molecule and a probe ion. In the present study, proton is introduced as an ion probe. the "ion probe"-"target molecule interaction. This method is applied to investigate the shape of H_2 O. We have obtained a closed surface of H_2O corresponding to the oxygen lone pairs. Shape similarity of deformed water molecules are numerically evaluated by referring to the calculated isovalue surface, Another application is carried out to glucose and psicose in order to see correlation with their biological activity.

View full abstract

We are developing a method to suggest functional materials by using the information obtained from electronic -structure calculations.In the present study, weapply a statistical analysis method in order to suggest a new tyrosinase inhibitor. We investigate a correlation between each descriptors about the electronic structure of the tyrosinase inhibitorand inhibitory activity. We performed a multivariate analysis using all descriptors and derived a predictive model of the inhibitory activity. To suggest a new candidate of a tyrosinase inhibitor, we investigate chemicals contained in crops by using a statistical predictive model. Some molecules are predicted to potentially have high activity by the present model, indicating that the present statistical analysis would be effective for predicting inhibitory activity.

View full abstract

Development of fluorescent bio-imaging molecules attracts much interest for in-vivo medicinal analysis. One of the typical examples is fluoroscein derivates called Tokyo Green (TG). We aim at developing an analytical method for data mining which would be useful for molecular design and discovery in Chemistry and Pharmaceutical Sciences. We have developed a computer program that automatically discovers a mathematical correlation function between an electronic descriptor and a chemical property. Such a correlation is revealed through a variety of fittings of experimentally measured values for various electronic factors such as orbital energy levels, ionization energies, reorganization energies, and so on. In this presentation, an application of the above computational method will be discussed for an organic pigment for bio-imaging. Herein we focus on TG. A large number of function fittings, it has been shown that the best fitting is obtained when the orbital energy of the molecular orbital localized on the benzene moiety is chosen as a descriptor.

View full abstract