We predicted the biodegradability of 554 chemicals by using a nonlinear partial least squares (PLS) method, called kernel PLS (KPLS), and compared the predictive performance of KPLS and that of linear PLS, which is widely used for modeling structure-activity relationships. Moreover, prediction using support vector machine (SVM) was performed to confirm the utility of KPLS. The KPLS models correctly categorized 429 (77.4%), 443 (80.0%) and 454 (81.9%) chemicals out of 554, whereas the PLS models were correct for 419 (75.6%), 434 (78.3%) and 439 (79.2%) in cases of using 6, 50 and 89 descriptors, respectively, based on the chemical structures of chemicals analyzed in this study. By properly tuning the necessary parameters, KPLS showed better predictive performance for the biodegradability of chemicals than PLS and SVM did, which showed 79.6% accuracy with 89 descriptors.

View full abstract

Quantitative Structure-Activity Relationship (QSAR) has been widely used in molecular design and many successful applications have been known. However, when precise model would be made, its chemical interpretation becomes difficult. Therefore, we have developed inverse QSAR system to generate chemical structures with high inhibitory activities. EA-Inventor (Evolutionary Algorithm-Inventor) was used as structure generator in our system. In EA-Inventor, initial structures represented by SMILES strings are modified using cross-over and mutation operations. The inhibitory activities are predicted by QSAR model and their values are used as the scores. The structures with high scores survive and next new structures are generated. After these cycles, the higher scored structures can be obtained. We have applied EA-Inventor to Trypsin inhibitory data. CoMFA (comparative molecular field analysis) is employed as QSAR model. If new structures differ from training set, high penalty values are added to scores in order to guarantee them within chemical space. The generated structures are high complementarily to Trypsin structure and the result seems to be reasonable. This technology can be extended to ADME model and chemical feasibility filter.

View full abstract

The foundation of the Cahn-Ingold-Prelog (CIP) system has been redefined by employing the concept of RS-stereogenicity derived from stereoisograms for analyzing stereoisomerism, where R- and S-stereodescriptors are pairwise given to two molecules in an RS-diastereomeric relationship. The conventional definitions of RS-stereodescriptors which are based on the original term "chirality" (along with enantiomeric and diastereomeric relationships) and on the revised term "stereogenicity" are altogether abandoned. Thereby, the present foundation is shown to be distinct from and to be more consistent than the conventional one which assigns R- and S-stereodescriptors to two molecules in an enantiomeric relationship as well as in a diastereomeric relationship. In order to harmonize the present foundation with the conventional one and to leave conventional results unchanged as far as possible, RS-stereodescriptors given originally to a pair of RS-diastereomers are considered to be translated into RS-stereodescriptors for a pair of enantiomers, where the concept of chirality-faithfulness of priority sequences has been proposed to testify whether the translation process of RS-stereodescriptors is faithful or not. As a result, the specification of configurations is rationalized on a more succinct basis than the conventional foundation which emphasizes the dichotomy between enantiomers and diastereomers. A paradigm shift from enantiomers to RS-stereoisomers has been pointed out so that over-simplified features of the conventional dichotomy between enantiomers and diastereomers have been discussed by means of stereoisograms.

View full abstract

A methodology for solving inverse-QSPR problem on designing novel polymers has been developed by chemoinfomatics approach composed of constructing predictive models with experimental data and searching for candidates which satisfy target properties by reverse analysis of the model. Through constructing Tg models of various homopolymers, the suitable size of chemical structure was figured out in modeling by group contribution method and RDF descriptor. These methods were found out to be useful for predicting Tg and other properties of random co-polyesters, which are combinations of diacids and diols. Using these models, multi-objective reverse analysis was carried out to find candidates which satisfy desired properties. The candidates were obtained by optimizing not only monomer compositions but also novel monomer structures. The agreement between prediction and experiment of some candidates confirmed that the proposed methodology is useful for polymer design.

View full abstract

We propose structural features of genomic DNA, which are essential to generate and to analyze genome. We calculated the appearance frequency of the nucleotides (bases) of throughout the entire genome as a polynucleotide molecule consisting of Adenine (A), Thymine (T), Guanine (G) and Cytosine (C) bases including the coding- and the non-coding regions, primarily in the genomes of Saccharomyces cerevisiae, Escherichia coli, and Homo sapiens. Our results indicate that the base sequences in a single-strand of DNA have the following characteristics: (1) reverse-complement symmetry of 3-9 successive bases, (2) bias and (3) multiple fractality of the distribution of four bases, A, T, G and C depending on the distance, exponentially decreased at short distances and linearly decreased at long distances in double logarithmic plot (power spectrum) of L (the distance of a base to the next base) vs P (L) (the probability of the base-distribution at L). These structural features of a single-strand of DNA can be clearly observed in any genomic DNA, especially observed remarkable in eukaryotic genome. Whereas in the artificial genomes or chromosomes with the same base-numbers, the same base-contents and the same frequencies of 64 triplets, the bias and the linearly-decreased fractality of the distribution of four bases described the above were missing, although the reverse-complement symmetry of the base sequences and the exponentially decreased-fractality of the base distribution were observed.

View full abstract

Precision agriculture is an agricultural concept doing the right things, in the right place, at the right time to increase production, decrease usage of fertilizer, pesticide, and water or protect the environment. It is necessary to understand accurate information of fields for proper field management. For this purpose, soil sensors have been studied to obtain soil information by predicting the amount of moisture and nitrogen in soil by using VIS-NIR spectrum. In this article, we built regression models between VIS-NIR spectrum and the amount of soil moisture, carbon, and nitrogen, electric conductivity, and pH by using PLS method. In addition, we proposed variable selection method, GAWLS, and validated availability and general versatility of the method. As a result, we revealed GAWLS is useful method for prediction of soil properties compared with PLS with all variables and GAPLS method.

View full abstract

Aryl hydrocarbon receptor (AhR) is a protein which induces expression of metabolic enzymes. Experimental studies indicated that biological functions of AhR are largely dependent on the organismic species. However, the structure of AhR and the recognition mechanism between AhR and ligands are not clarified yet. In the present study, we investigated specific interactions between AhR of beluga and tern and TCDD (2,3,7,8-tetrachrodibenzo-p-dioxin) by using fragment molecular orbital calculations. The results indicate that the binding energy between beluga AhR and TCDD is larger than that between tern AhR and TCDD, which is consistent with the experimental result on the affinity between TCDD and these AhRs. In addition, we clarified the important amino acid residues of AhRs for binding TCDD.

View full abstract

The pro-R/pro-S system has been redefined by using the concept of pro-RS-stereogenicity, where two ligands in an RS-diastereotopic relationship are pairwise characterized by pro-R- and pro-S-descriptors on the basis of stereoisograms. As for a practical criterion for determining RS-diastereotopic relationships, a symmetry criterion has been developed as a new matter by extending the concept of stereoisograms to testify RS-diastereotopic relationships. The conventional definitions of pro-R/pro-S-descriptors which are based on the original term "prochirality" (along with enantiotopic and diastereotopic relationships) and on the revised term "prostereogenicity" (along with stereoheterotopic relationships) are altogether abandoned. The revised concept of prochirality has a purely geometric meaning, which stems from the term enantiotopic relationship or equivalently enantiospheric equivalence class (S. Fujita, Symmetry and Combinatorial Enumeration in Chemistry, Springer-Verlag, Berlin-Heidelberg (1991)).The two concepts for describing intramolecular environments, i.e., the pro-RS-stereogenicity and the prochirality, are clarified to correspond to the concepts of RS-stereogenicity and chirality, which have been formulated on the basis of stereoisograms in order to bring about a harmonized viewpoint to stereoisomeric and geometric intermolecular features of stereochemistry (S. Fujita, J. Org. Chem., 69, 3158-3165 (2004); J. Comput. Aided Chem., 10, 16-29 (2009)). Chirality-faithfulness of pro-R/pro-S-descriptors specified by RS-diastereotopic relationships is discussed to harmonize pro-RS-stereogenicity with prochirality.

View full abstract

CO₂ capture and storage (CCS), which is one of technologies against global warming, attracts global attention. However, cost of CCS would be very high, and it is hard to put it into practical use. In particular, capturing CO₂ represent a major proportion of overall CCS cost. In this paper, we noted capturing CO₂ by the chemical absorption method using alkanol amines and we aimed to research new alkanol amine. Alkanol amine-water solutions react with CO₂, forming chemical compounds that separate CO₂ from the gas mixtures at a higher rate than the natural CO₂ absorption in pure water. Alkanol amine can be recovered by diffusion of CO₂. For decreasing the cost of CO₂ capture, we need new alkanol amines to react quickly with CO₂ and to require less heat of reaction. So we aimed to research alkanol amines which meet the requirement of them. Regression models to predict CO₂ absorption rate and heat of reaction were constructed by PLS (partial least squares) method and GAPLS (genetic algorithm based-PLS) method. As a result, R² and Q² values of the GAPLS of heat of reaction are 0.999 and 0.990, respectively. These of CO₂ absorption rate are 0.957 and 0.914. Thus these models are expected to have high predictive accuracy. Then these models were used for selecting alkanol amines expected to be efficient absorbent from chemical structure which is virtually generated in computer. As a result, we got several kinds of new alkanol amines to be more efficient absorbent than existing ones.

View full abstract

A new capability has been added to the synthesis design system SYNSUP [1]. The new module generates synthetic routes that focus on key-bonds. Key-bonds are selected according to their spatial relationships to chiral atoms and central atoms of functional groups. When the key-bonds cannot be formed by any reaction known to SYNSUP, then structure changes in the vicinity of the bonds are effected so that the key-bonds can be disconnected in a retro-synthetic pathway. We called the fashion “selective search”. Such structure changes were studied, using 27 reported total synthesis of 25 naturally occurring compounds. This kind of selective search has been shown to be useful and practical for relatively complicated molecules.

View full abstract

The adverse events induced by drugs have been complicated, when two or more drugs are administrated for a patient. We selected "Stevens-Johnson Syndrome (SJS) " as a research object, which is one of the severe skin manifestations. The data source is a database constructed by the Food and Drug Administration (FDA). FDA's post-marketing safety surveillance program is supported by the Adverse Event Reporting System (AERS). AERS is designed with a computerized information database. To analyze the relationships between the concurrent medication and SJS in this study, we applied association rule learning. Our purpose is to propose an efficient procedure that enables the detection of signals for drugs related to an adverse event, without assuming the involvement of a specific drug. We defined new value K for the evaluation of existing signal detection. Association rule was evaluated according to criterion K value. As a result, it was suggested to obtain a strong signal by combining two concomitant drugs. Association rule learning in this study was applicable for the analysis of the relationships between adverse events and pairs of drugs.

View full abstract