Dissertations / Theses on the topic 'Cheminformatics and quantitative structure-activity relationships'

A good QSAR model comprises several components. Predictive accuracy is paramount, but it is not the only important aspect. In addition, one should apply robust and appropriate statistical tests to the models to assess their significance or the significance of any apparent improvements. The real impact of a QSAR, however, perhaps lies in its chemical insight and interpretation, an aspect which is often overlooked. This thesis covers three main topics: a comparison of contemporary classifiers, interpretability of random forests and usage of interpretable descriptors. The selection of data mining technique and descriptors entirely determine the available interpretation. Using interpretable approaches we have demonstrated their success on a variety of data sets. By using robust multiple comparison statistics with eight data sets we demonstrate that a random forest has comparable predictive accuracies to the de facto standard, support vector machine. A random forest is inherently more interpretable than support vector machine, due to the underlying tree construction. We can extract some chemical insight from the random forest. However, with additional tools further insight would be available. A decision tree is easier to interpret than a random forest. Therefore, to obtain useful interpretation from a random forest we have employed a selection of tools. This includes alternative representations of the trees using SMILES and SMARTS. Using existing methods we can compare and cluster the trees in this representation. Descriptor analysis and importance can be measured at the tree and forest level. Pathways in the trees can be compared and frequently occurring subgraphs identified. These tools have been built around the Weka machine learning workbench and are designed to allow further additions of new functionality. The interpretability of a model is dependent on the model and the descriptors. They must describe something meaningful. To this end we have used the TMACC descriptors in the Solubility Challenge and literature data sets. We report how our retrospective analysis confirms existing knowledge and how we identify novel C-domain inhibition of ACE. In order to test our hypotheses we extended and developed existing software forming two applications. The Nottingham Cheminformatics Workbench (NCW) will generate TMACC descriptors and allows the user to build and analyse models, including visualising the chemical interpretation. Forest Based Interpretation (FBI) provides various tools for interpretating a random forest model. Both applications are written in Java with full documentation and simple installations wizards are available for Windows, Linux and Mac.

The polychlorinated biphenyls (PCBs) comprise a group of 209 congeners varying in the number of chlorine atoms and substitution patterns. These compounds tend to be biomagnified in foodwebs and have been shown to induce an array of effects in exposed organisms. The structural characteristics of the PCBs influence their potency as well as mechanism of action. In order to assess the biological potency of these compounds a multi-step quantitative structure-activity relationship (QSAR) procedure was used in the project described in this thesis.

The ultraviolet absorption (UV) spectra were measured for all 209 PCBs, and digitised for use as physico-chemical descriptors. Interpretations of the spectra using principal component analysis (PCA) showed the number of ortho chlorine atoms and para-para substitution patterns to be significant. Additional physico-chemical descriptors were derived from semi-empirical calculations. These included various molecular energies, the ionisation potential, electron affinity, dipole moments, and the internal barrier of rotation. The internal barrier of rotation was especially useful for describing the conformation of the PCBs on a continuous scale.

In total 52 physico-chemical descriptors were compiled and analysed by PCA for the tetra- to hepta-chlorinated congeners. The structural variation within these compounds was condensed into four principal properties derived from a PCA for use as design variables in a statistical design to select congeners representative for these homologue-groups. The 20 selected PCBs have been applied to study structure-specific biochemical responses in a number of bioassays, and to study the biomagnification of the PCBs in various fish species.

QSARs were established using partial least squares projections to latent structures (PLS) for the PCBs potency to inhibit intercellular communication, activate respiratory burst, inhibit dopamine uptake in synaptic vesicles, compete with estradiol for binding to estrogen receptors, and induce cytochrome P4501A (CYP1A) related activities. By the systematic use of the designed set of PCBs the biological potency was screened over the chemical domain of the class of compounds. Further, sub-regions of highly potent PCBs were identified for each response measured. For risk assessment of the PCBs potency to induce dioxin-like activities the predicted induction potencies (PIPs) were calculated. In addition, two sets of PCBs were presented that specifically represent congeners of environmental relevance in combination with predicted potency to induce estrogenic and CYP1A related activities.

Quantitative Structure-Activity relationship (QSAR) attempt statistically to relate the physico-chemical properties of a molecule to its biological activity. A QSAR analysis was performed on the toxicities of up to 75 organic chemicals to two aquatic species, Photobacterium phospherum (known as the Microtox test), and the fathead minnow. To model the toxicities 49 physico-chemical and structural parameters were produced including measures of hydrophobicity, molecular size and electronic effects from techniques such as computational chemistry and the use of molecular connectivity indices. These were reduced to a statistically more manageable number by cluster analysis, principal component analysis, factor analysis, and canonical correlation analysis. The de-correlated data were then used to form relationships with the toxicities. All the techniques were validated using a testing set. Some good predictions of toxicity came from regression analysis of the original de-correlated variables. Although successful in simplifying the complex data matrix, principal component analysis, factor analysis, and canonical content analysis were disappointing as predictors of toxicity. The performance of each of the statistical techniques is discussed. The inter-species relationships of toxicity between four Commonly utilised aquatic endpoints, fathead minnow 96 hour IC50, Microtox 5 minute EC50, Daphnia magna 48 hour IC50, and Tetrahymena pyriformis 60 hour IG50, were investigated. Good relationships was found between the fathead minnow and both T. pyriformis and D. magna toxicities indicating that these species could be used to model fish toxicity. The outliers from individual relationships were assessed in order to elucidate if any molecular features may be causing greater relative toxicity in one species as compared to another. It is concluded that in addition to the intrinsic differences between species, the greater length of the test time for any species may result in increases bioaccumulation, metabolism, and detoxification of certain chemical classes. The relationships involving fish toxicity were moderately improved by the addition of a hydrophobic parameter.

Quantitative structure-activity relationships (QSAR) rationalize interrelation between molecular structure and biological response in terms of either physicochemical parameters, as in linear free energy relationships (LFER), or via purely empirical parameters, as is the case for De Novo schemes. In LFER the leading process is often the partitioning of a compound between two solvent phases, taken to represent the transfer of a drug molecule across a biological membrane. This study has investigated the partitioning behaviour of three series of hydroxybenzoate esters, viz. o-, m- and predominantly p-esters, the latter being preservatives in pharmaceutical formulations. The thermodynamic parameters AH, AG and AS for the transfer process were derived in an attempt to establish a QSAR. on a fundamental thermodynamic basis. Such parameters have identifiable physicochemical meaning and lend themselves more readily to interpretation. This facilitates application to alternative systems. A new Gibbs function factor analysis was developed and utilized to obtain thermodynamic contributions for parent and incremental methylene group portions of thestudy molecules. The empirical Collander equation for interrelation of various solute/solvent systems was also rationalized on a thermodynamic basis. Further extension of the Gibbs function factor analysis allowed scaling of "solvent" systems including chromatographic packings, solvents and liposomes. The scheme indicated capacity for optimized selection of bulk solvent systems to mimic biological membranes. A novel analytical procedure for direct measurement of biological response was developed. The bioassay appeared capable of discrimination i) between the closely related structural homologues, ii) between gram-negative and gram-positive bacteria, and further, iii) between certain cell batches of the same bacteria type. Also, the bioassay demonstrated a Collander interrelation between the two bacteria types. Flow microcalorimetry was the technique employed to measure thermal response of respiring E. coli and Staph, aur. bacteria. The modification of biological response with drug concentration was quantitated and a log dose max term was derived for each homologue. The results indicated potential for a predictive, additive structure-activity scheme based on assessment of biological response (BR) direct rather than through f(BR) via physicochemical or empirical parameters.

A Quantitative Structure-Activity Relationship (QSAR) study is an attempt to model some biological activity over a collection of chemical compounds in terms of their structural properties A QSAR model may be constructed through (typically linear) multivariate regression analysis of the biological activity data against a number of features or 'descriptors' of chemical structure. As with any regression model, there are a number of issues emerging in real applications, including (a) domain of applicability of the model, (b) validation of the model within its domain of applicability, and (c) possible non-linearity of the QSAR Unfortunately the existing methods commonly used in QSAR for overcoming these issues all suffer from problems such as computational inefficiency and poor treatment of non- linearity. In practice this often results in the omission of proper analysis of them altogether. In this thesis we develop methods for tackling the issues listed above using K-means clustering. Specifically, we model the shape of a dataset in terms of intelligent K-means clustering results and use this to develop a non- parametric estimate for the domain of applicability of a QSAR model. Next we propose a 'hybrid' variant of K-means, incorporating a regression-wise element, which engenders a technique for non-linear QSAR modelling. Finally we demonstrate how to partition a dataset into training and testing subsets, using the K-means clustering to ensure that the partitioning respects the overall distribution Our experiments involving real QSAR data confirm the effectiveness of the methods developed in the project.

Studies into the quantitative structure act! Vlty relationships of rate s of I-methyl-4-phenyl-l,2,3,6-tetrahydropyridine oxidation catalyzed by monoamine oxidases A and B were performed to elucidate active site substrate conformation and oxidation mechanisms. Plotting experimental kinetic activity against molecular properties obtained by experiment and by computational chemistry methods demonstrated correlations with lipophilic, steric, and electronic factors. Compounds studied were 4-aryloxy analogs, 4- aromatic heterocycle analogs, and 4-phenyl analogs. The conformer with phenyl ring to tetrahydropyridine dihedral angles similar to a low energy conformer of I-methyl-4-(2'-methyl-phenyl)-1,2,3,6- tetrahydropyridine is the most active conformer. Results indicate that rate limiting single electron transfer mechanisms are more viable than hydrogen atom abstraction mechanisms. Results indicate that binding or dissociation is the rate limiting step for aryloxy-analog oxidation catalyzed by monoamine oxidase B whereas the catalytic event itself is the rate limiting step for the other analogs. Several equations were developed to describe quantitative structure activity relationships of oxidation rates.
Master of Science

Preparou-se, neste trabalho, nove cloridratos de N,N- [(dimetilamino)metil]benzamidas-p-X-substituidas em que X = N02, Br, Cl, I, F, H, CH3, OCH3 e N(CH3)2, SÉRIE II, ainda não descritos na literatura, a partir das respectivas bases, SÉRIE I. Os compostos obtidos foram identificados por seus espectros de I.V e RMN1H, e suas purezas determinadas por análise elementar e ponto de fusão. A seguir, determinou-se os valores de vC=O (cm-l ) em HCCl3, como medida de sua polaridade tanto para os compostos da SÉRIE I como para os da SÉRIE II. A aplicação da equação de HAMMETT, simples e expandida, aos valores obtidos forneceu excelentes correlações, verificando-se que os efeitos eletrônicos que os substituintes exercem sobre o grupo carbonila são de natureza indutiva e de ressonância, não havendo predominância de um sobre o outro. Determinou-se também, por bloqueio nervoso periférico em pata de rato, o grau de atividade anestésica local de sete dos nove compostos da SÉRIE II. A aplicação da equação de HANSCH aos valores de atividade anestésica local mostrou haver correlação razoável considerando-se os efeitos eletrônicos e hidrofóbicos exercidos pelos substituintes. A análise dos coeficientes de regressão obtidos sugere haver contribuição preponderante dos efeitos eletrônicos em relação à contribuição do efeito hidrofóbico para a atividade anestésica local.
In the present work nine N,N-[(dimethyl amino)methyl]benzamides-p-X-substituededs hydrochlorides (SERIES II) were prepared from the corresponding bases, (SERIES I), where X = N02, Br, Cl, I, F, H, CH3, OCH3 and N(CH3)3. The purity of the novel compounds, SERIES II, was established by elemental analysis. The recorded infrared and 1HNMR spectra were in agreement with their structures. For both sets of compounds the carbonyl group infrared frequencies in HCCl3 were determined and used as measurement of the polarity of the group. The obtained vC=O (cm-1) values showed excellent correlations when a simple or a multiparameter HAMMETT equation was applied. This suggests that the eletronic effect of the substituents on the carbonyl stretching frequencies is of inductive and ressonanee nature, without predominances of one over the other. Were also determined by peripheryc nervous blockade on rats paw, the degree of local anesthetic activity on seven of the nine compounds of SERIES II. The application of the HANSCH equation to the local anesthetie activity values showed reasonable correlation when the eletronic and hydrophobic effects of the substituents were considered. Analysis of the obtained regression coefficents suggest that the contribution of the former effect to the local anesthetic activity predominants.

Bases de Mannich têm sido sintetizadas como pró-fármacos de cetonas α,β-insaturadas, 22, 20 sendo estas importantes no tratamento do câncer. 20 Assim, toma-se de interesse o desenvolvimento de estudos de QSAR envolvendo bases de Mannich com propriedades citotóxicas e/ou anticâncer, contribuindo-se, para o entendimento da(s) interação(ões) destes compostos (agentes alquilantes) no sistema biológico. 40, 5, 36, 53 Neste trabalho, destinado a dissertação de mestrado, foram preparadas e purificadas por métodos triviais descritos na literatura 8 47 duas séries de derivados de bases de Mannich, a saber: nove cloretos de 3-(dimetilamino)-propiofenonas-4-X-substituidas (série I, compostos I.1 a I.9), e seis dos correspondentes iodetos de 3-(trimetilamino )propiofenonas-4-X-substituídas (série II, compostos II.1 a II.6). Destes, quatro não estão descritos na literatura. A seleção dos substituintes foi feita visando obter intercorrelações não significativas entre os correspondentes valores dos parâmetros físico-químicos analisados (π, σp e MR4). 45 Outro critério considerado, se refere as faixas de variação para cada parâmetro físico-químico analisado. 62, 31, 15 Para a série II, estes critérios foram parcialmente contemplados. Em seguida, para cada composto preparado, foram determinados experimentalmente e/ou retirados da literatura 45 e/ou calculados parâmetros parâmetros fisico-químicos/estruturais descritores de propriedades, respectivamente: hidrofóbicas/lipofílicas (π, logPcalc e logP app (sendo os valores este último determinados somente para os compostos da serie I)); eletrônicas/polares (σp, σp+, σp-, σI, σR, T, R, γ13C=0 e νC=0) e, ainda aquela relativa à polarizabilidade (MR4). Como parâmetro biológico para os compostos das series I e II foram determinados os valores da potência citotóxica, expressos por log(1/IC50). A seguir, com o objetivo de se desenvolver estudos de QSAR, aplicando a abordagem tradicional 40 para os compostos respectivamente das séries I e II, e a abordagem mista, 62, 63, 53 que considera as duas séries conjuntamente, foram propostos respectivamente modelos expressos pelas equações IV.4 e IV.6: log(1/IC50)= -0,27(±0,23)δ13C=0 -0,48(±0,35)logPcalc + 56,32(±44,30) eq.IV.4 (n=9;r=0,87;s=0,17;F=9,23;Q2=0,36;SPRESS=0,28) log(1/IC50)=-0,56(±0,27)π+0,35(±0,29)MR4+0,23(±0,22)I[N(Me)3]+-I+1,53(±0,22) eq.IV.6 (n=15;r=0,82;s=0,18;F=7,64;Q2=0,43;SPRESS=0,24 Primeiramente, através da análise do modelo expresso pela equação IV.4 foi possível avaliar as naturezas e contribuições relativas dos parâmetros estruturais responsáveis pela citotoxicidade dos compostos da série I. Em seguida, através da análise do modelo expresso pela equação IV.6 foi possível avaliar a contribuição dos parâmetros físico-químicos, bem como a contribuição da variação estrutural existente entre as séries I e II, responsáveis pela citotoxicidade dos compostos. A interpretação e significado de I[N(Me)3]+I-, que assume valor 0 para os compostos da série I e valor 1 para os compostos da série II, foi discutida em termos estruturais. Nenhum modelo com significado estatístico foi obtido para os compostos da série II. E, tanto para a série I, bem como para as séries I e II, a aplicação dos modelos tanto parabólico como bilinear não resultou em correlações estatisticamente significativas. Pela análise dos modelos foi possível, pelo menos em parte, o entendimento da(s) interação(ões) dos derivados de bases de Mannich no sistema biológico. 40, 5, 36, 53.
Mannich bases have been used as prodrugs of α,β-unsaturated ketones that are important in the cancer therapy 80, 22. In this way, a QSAR study 20, 81 was performed with two sets of Mannich bases derivatives, namely: 3-(dimethylamine)-propiophenon-4-X-substituted hydrochlorides (set I, composed of nine derivatives) and 3-(trimethylamine)propiophenon-4-X-substituted iodines (set II, composed of six derivatives). The sets I and II were prepared by trivial methods described in the literature 58,8. For each compound the physicochemical/structural descriptors, hydrophobic/lipophilic (π, logPcalc and logP app (it was determined only for set I), electronic/polar (σp, σp+, σp-, σI, σR, T, R, γ13C=0 e νC=0) and the polarizability related (MR4), were determined experimentally and/or calculated and/or obtained from the literature 7. The biological parameter was the cytotoxic potency, expressed by values of log(1/IC50). In order to investigate 1he interaction of this class of compounds wi1h the biological system a QSAR study was performed 20, 81. The most significant QSAR models obtained for set I and sets I and II altogether, were expressed respectively by equations 1 and 2. log(1/IC50)= -0,27(±0,23)δ13C=0 -0,48(±0,35)logPcalc + 56,32(±44,30) eq.1 (n=9;r=0,87;s=0,17;F=9,23;Q2=0,36;SPRESS=0,28) log(1/IC50)=-0,56(±0,27)π+0,35(±0,29)MR4+0,23(±0,22)I[N(Me)3]+-I+1,53(±0,22) eq.2 (n=15;r=0,82;s=0,18;F=7,64;Q2=0,43;SPRESS=0,24 To equation 2 was included an variable indicator I[N(Me)3]+I- that assumed the value 0 for set I compounds and the value of 1 for set II compounds. The interpretation and meaning of I[N(Me)3]+I-, were discussed in structural terms. No significant models were obtained for the compounds of the set II. For set I and sets I and II altogether, the application of the parabolic and the bilinear models were verified and showed to be not statistically significant.

O estudo das Relações Quantitativas entre Estrutura química e Atividade biológica (QSAR/QSAR-3D) utilizando diferentes estratégias metodológicas complementares, aplicadas iterativamente, se estende a diferentes áreas de aplicação. Dentre elas destaca-se o planejamento racional de novos fármacos e o estudo de seus mecanismos de ação. Moléculas candidatas a fármacos, geradas em estudos de QSAR-3D, freqüentemente, não se tomam fármacos por apresentarem inadequadas propriedades farmacocinéticas, ou seja, ADME (absorção, distribuição, metabolismo e excreção). Nota-se assim que o conhecimento das propriedades farmacocinéticas de ligantes deve ser uma preocupação presente, em qualquer nível de desenvolvimento de um novo fármaco e ressalta-se a importância do QSAR como uma ferramenta de grande valor, principalmente na compreensão da contribuição das propriedades físico-químicas, identificadas como sendo responsáveis pela atividade. A aplicação do QSAR permite elucidar a natureza e a grandeza das propriedades físico-químicas e estruturais nas interações entre o composto e o seu alvo biológico nos processos farmacocinéticos de ADME (TESTA, 2001; MALVEZZI et al.,, 2001; GAVIRAGHI et al., 2001). Recentemente, em trabalhos realizados em nosso laboratório (SIQUEIRA, 2001), envolvendo a análise de QSAR aplicada aos valores do bloqueio da transmissão neuromuscular determinados para uma série de onze brometos de [2-(4-X-benzamido)etil]benzildimetilamônio substituídos, demonstrou-se tanto a importância da presença de um grupo catiônico terminal como da lipofilicidade para o bloqueio da transmissão neuromuscular. No presente trabalho, visando descrever as interações hidrofóbicas determinantes da potencial atividade bloqueadora neuromuscular, para uma série de dez derivados da procaína, sendo eles: cloretos de 3-(dimetilamino)propiofenona (composto 1); de 3-(dietilamino)propiofenona (composto 2); de 3-(1-pirrolidino)propiofenona (composto 3); de 3-(1-piperidino)propiofenona (composto 4); de 3-(1-morfolino)propiofenona (composto 5); de 3-[1-(4-metilpiperazino)] propiofenona (composto 6); de 3-(c-hexilamino)propiofenona (composto 7); de N-[(N,N-dimetilamino)etil]benzamida (composto 8); de N-[(N,N-dimetilamino )metil]benzamida (composto 9) e, de N-[(N,N-dimetilamino)etil]benzoato (composto 10), por nós sintetizados, determinaram-se os respectivos parâmetros hidrofóbicos e relacionados à ionização (logPappSF; logPnSF; logPiSF; (pKaaq)SF; (pKaoct)SF; logPappHPLC; IogPnPot; IogPiPot; (pKaaq)Pot; logPcalc) obtidos por três métodos experimentais (por Shake-flask; por HPLC e potenciométrico) e por cálculo (programa CLOGP), como modelos de partição. Através da análise comparativa dos valores dos parâmetros hidrofóbicos e relacionados à ionização para a série de dez derivados da procaína (compostos 1-10) sintetizados neste trabalho foi possível constatar que: (i) os valores de IogPcalc modelam, apenas, parcialmente as interações hidrofóbicas refletidas nos valores de logPappSF e de logPappHPLC em valores de pH 3,0 e 10,0, enquanto que as interações hidrofóbicas refletidas nos valores de logPnPot são idealmente modeladas pelos valores de logPcalc; (ii) nas mesmas condições experimentais, os métodos Shake-flask, HPLC e potenciométrico revelam interações hidrofóbicas proporcionais, entretanto, correlações não ideais foram observadas quanto comparados entre si; (iii) os valores de pKaSF , determinados indiretamente por Shake-flask, foram validados pela determinação direta de pKaPot, por potenciometria. Embora o sistema de avaliação da bloqueadora da transmissão neuromuscular tenha sido adequado, a série de compostos estudada não gerou valores válidos para a construção das correspondentes curvas dose-resposta. Os valores obtidos de lC50 para os compostos 3, 4 e 8 da série estudada foram insuficientes, em número, para serem utilizados em uma análise de QSAR. Estudos futuros podem ser realizados com o objetivo de verificar a atividade anestésica local para os compostos da série estudada, avaliando-se a perda da sensibilidade à estimulação sensorial, em cobaias, ou em estudos mais sofisticados de eletrofisiologia (patch clamp). Poder-se-á, assim, fornecer subsídios para elucidar as contribuições relativas da forma catiônica e da forma neutra, respectivamente de compostos ionizáveis com atividade anestésica local.
The scope of Quantitative Structure and Biological Activity Relationships (QSAR/QSAR-3D) studies, using distinct complementary methods, spreads over multiple areas of applicability. Among those, striking are the rational design of new drugs and the enlightenment of their mechanism of action. Candidate molecules, shaped by QSAR-3D analysis, frequently fail to become drugs, and hit the market, due to insufficient pharmacokinetic properties (i.e. absorption, distribution, metabolism and excretion; ADME). Hence, the awareness of pharmacokinetic properties of molecules should be a concern at any leveI of development of new drug candidates. Noteworthy is the use of QSAR as a tool of great value, specially, in the comprehension of the contribution of the physical-chemical properties, responsible for the biological activity. (TESTA, 2001; MALVEZZl et ai., 2001; GAVlRAGHl et ai., 2001). ln this work, aiming to describe the hydrophobic interactions determining the neuromuscular blocking activity of a series of ten procaine derivatives: the hydrochlorides of 3-(dimethylamine)propiophenone (compound 1); of 3-(diethylamine)propiophenone (compound 2); of 3-(1-pirrolidine)propiophenone (compound 3); of 3-(1-piperidine)propiophenone (compound 4); of 3-(1-morfoline)propiophenone (compound 5); of 3-[1-(4-methylpiperazine)]propiophenone (compound 6); de 3-(c-hexilamine)propiophenone (compound 7); de N-[(N,N-dimethylamine)ethyl]benzamide (compound 8); de N-[(N,N-dimethylamine)methyl]benzamide (compound 9) and, of N-[(N,N-dimethylamine)ethyl]benzoate (compound 10), prepared by us, the corresponding hydrophobic parameters (logPappSF; logPnSF; logPiSF; (pKaaq)SF; (pKaoct)SF; logPappHPLC; IogPnPot; IogPiPot; (pKaaq)Pot; logPcalc) were obtained by three experimental methods (Shake-flask; HPLC and potenciometric) and by calculus (CLOGP) as models of partition and distribution. Following the comparative analysis of the values of the hydrophobic parameters determined for the ten compounds of the series of procaine analogs, it was possible to conclude that: (i) the values of IogPcalc reflect only partially, the hydrophobic interactions encoded in the values of logPappSF and logPappHPLC determined in pH values of 3,0 and 10,0; (ii) the values of logPcalc reflect the hydrophobic interactions encoded in the values of logPnPot; (iii) in the same experimental conditions, the methods Shake-flask, HPLC and potenciometric, reflect proportional hydrophobic interactions, although the observed correlations were not ideal when the methods were compared between each other; (iv) the values of pKaSF, determined indirectly by Shake-flask, were validated by the direct measurement of pKaPot, by potentiometry. Even though the system used to measure the neuromuscular blocking actions was adequate, the series of procaine derivatives failed to provide sufficient values to prepare dose-effect curves and proceed with the quantitative structure activity relationships.

En esta tesis se ha utilizado la metodología QSPR para calcular las propiedades de diferentes compuestos y sistemas complejos que no habían estudiados anteriormente. En concreto, se han establecido modelos que permiten el cálculo de la viscosidad y la tensión superficial, en estado líquido, y la entalpía de formación en fase gas para conjuntos de compuestos organometálicos de fórmula general MRnXm, en la que M puede ser un metal, semimetal o no metal de los grupos 12 al 16 de la tabla periódica; los grupos R corresponden a sustituyentes orgánicos alquílicos, arílicos, etc.; y los ligandos terminales X pueden ser cloro, bromo, yodo e hidrógeno. Se ha estudiado también la basicidad catiónica de un conjunto de compuestos orgánicos, de naturaleza química muy diversa, frente al catión Li+. En general, esta propiedad puede asociarse a la energía del proceso de formación de los complejos [Li-Ligando]+. Los sistemas complejos estudiados, que reciben el nombre de multicomponentes, son aquellos en los que la propiedad estudiada depende, a la vez, de dos o más elementos constituyentes del sistema. Las propiedades estudiadas en esta tesis son: las afinidades y basicidades catiónicas de los aminoácidos habituales frente a los cationes monovalentes de litio, sodio, potasio, cobre y plata; y las constantes de acidez (pKa) de familias de ácidos orgánicos en diferentes solventes, en este caso las familias de ácidos orgánicos estudiadas son fenoles, ácidos benzoicos, ácidos carboxílicos alifáticos y anilinas. En el tratamiento de estos sistemas multicomponentes se han utilizado descriptores externos para caracterizar a los cationes metálicos y los solventes. En el primer caso se han utilizado propiedades físicas, como potenciales de ionización, afinidades electrónicas, escalas de electronegatividad, etc.; para los diferentes solventes se han usado también propiedades físicas, como el momento dipolar, la constante dieléctrica, la polarizabilidad, etc.; y parámetros derivados de las diferentes escalas de polaridad más habituales, como los parámetros de Kamlet y Taft, los de Koppel y Palm, los de Drago, Gutmann, etc. Los modelos, lineales y no lineales, desarrollados para todas las propiedades proporcionan resultados excelentes para todas ellas, con valores de R2 mayores que 0.95, errores muy bajos y capacidad predictiva elevada, comprobada mediante la utilización de conjuntos de valores externos. Además de proporcionar una manera de calcular las propiedades, los modelos establecidos contienen descriptores que permiten realizar, en todos los casos, una interpretación razonable de las propiedades en términos fisicoquímicos.

La present tesi, tot i que emmarcada dins de la teoria de les Mesures Semblança Molecular Quántica (MQSM), es deriva en tres àmbits clarament definits:
- La creació de Contorns Moleculars de IsoDensitat Electrònica (MIDCOs, de l'anglès Molecular IsoDensity COntours) a partir de densitats electròniques ajustades.
- El desenvolupament d'un mètode de sobreposició molecular, alternatiu a la regla de la màxima semblança.
- Relacions Quantitatives Estructura-Activitat (QSAR, de l'anglès Quantitative Structure-Activity Relationships).
L'objectiu en el camp dels MIDCOs és l'aplicació de funcions densitat ajustades, ideades inicialment per a abaratir els càlculs de MQSM, per a l'obtenció de MIDCOs. Així, es realitza un estudi gràfic comparatiu entre diferents funcions densitat ajustades a diferents bases amb densitats obtingudes de càlculs duts a terme a nivells ab initio. D'aquesta manera, l'analogia visual entre les funcions ajustades i les ab initio obtinguda en el ventall de representacions de densitat obtingudes, i juntament amb els valors de les mesures de semblança obtinguts prèviament, totalment comparables, fonamenta l'ús d'aquestes funcions ajustades.
Més enllà del propòsit inicial, es van realitzar dos estudis complementaris a la simple representació de densitats, i són l'anàlisi de curvatura i l'extensió a macromolècules. La primera observació correspon a comprovar no només la semblança dels MIDCOs, sinó la coherència del seu comportament a nivell de curvatura, podent-se així observar punts d'inflexió en la representació de densitats i veure gràficament aquelles zones on la densitat és còncava o convexa. Aquest primer estudi revela que tant les densitats ajustades com les calculades a nivell ab initio es comporten de manera totalment anàloga. En la segona part d'aquest treball es va poder estendre el mètode a molècules més grans, de fins uns 2500 àtoms.
Finalment, s'aplica part de la filosofia del MEDLA. Sabent que la densitat electrònica decau ràpidament al allunyar-se dels nuclis, el càlcul d'aquesta pot ser obviat a distàncies grans d'aquests. D'aquesta manera es va proposar particionar l'espai, i calcular tan sols les funcions ajustades de cada àtom tan sols en una regió petita, envoltant l'àtom en qüestió. Duent a terme aquest procés, es disminueix el temps de càlcul i el procés esdevé lineal amb nombre d'àtoms presents en la molècula tractada.
En el tema dedicat a la sobreposició molecular es tracta la creació d'un algorisme, així com la seva implementació en forma de programa, batejat Topo-Geometrical Superposition Algorithm (TGSA), d'un mètode que proporcionés aquells alineaments que coincideixen amb la intuïció química. El resultat és un programa informàtic, codificat en Fortran 90, el qual alinea les molècules per parelles considerant tan sols nombres i distàncies atòmiques. La total absència de paràmetres teòrics permet desenvolupar un mètode de sobreposició molecular general, que proporcioni una sobreposició intuïtiva, i també de forma rellevant, de manera ràpida i amb poca intervenció de l'usuari. L'ús màxim del TGSA s'ha dedicat a calcular semblances per al seu ús posterior en QSAR, les quals majoritàriament no corresponen al valor que s'obtindria d'emprar la regla de la màxima semblança, sobretot si hi ha àtoms pesats en joc.
Finalment, en l'últim tema, dedicat a la Semblança Quàntica en el marc del QSAR, es tracten tres aspectes diferents:
- Ús de matrius de semblança. Aquí intervé l'anomenada matriu de semblança, calculada a partir de les semblances per parelles d'entre un conjunt de molècules. Aquesta matriu és emprada posteriorment, degudament tractada, com a font de descriptors moleculars per a estudis QSAR. Dins d'aquest àmbit s'han fet diversos estudis de correlació d'interès farmacològic, toxicològic, així com de diverses propietats físiques.
- Aplicació de l'energia d'interacció electró-electró, assimilat com a una forma d'autosemblança. Aquesta modesta contribució consisteix breument en prendre el valor d'aquesta magnitud, i per analogia amb la notació de l'autosemblança molecular quàntica, assimilar-la com a cas particular de d'aquesta mesura. Aquesta energia d'interacció s'obté fàcilment a partir de programari mecanoquàntic, i esdevé ideal per a fer un primer estudi preliminar de correlació, on s'utilitza aquesta magnitud com a únic descriptor.
- Càlcul d'autosemblances, on la densitat ha estat modificada per a augmentar el paper d'un substituent. Treballs previs amb densitats de fragments, tot i donar molt bons resultats, manquen de cert rigor conceptual en aïllar un fragment, suposadament responsable de l'activitat molecular, de la totalitat de l'estructura molecular, tot i que les densitats associades a aquest fragment ja difereixen degut a pertànyer a esquelets amb diferents substitucions. Un procediment per a omplir aquest buit que deixa la simple separació del fragment, considerant així la totalitat de la molècula (calcular-ne l'autosemblança), però evitant al mateix temps valors d'autosemblança no desitjats provocats per àtoms pesats, és l'ús de densitats de Forats de fermi, els quals es troben definits al voltant del fragment d'interès. Aquest procediment modifica la densitat de manera que es troba majoritàriament concentrada a la regió d'interès, però alhora permet obtenir una funció densitat, la qual es comporta matemàticament igual que la densitat electrònica regular, podent-se així incorporar dins del marc de la semblança molecular. Les autosemblances calculades amb aquesta metodologia han portat a bones correlacions amb àcids aromàtics substituïts, podent així donar una explicació al seu comportament.
Des d'un altre punt de vista, també s'han fet contribucions conceptuals. S'ha implementat una nova mesura de semblança, la d'energia cinètica, la qual consisteix en prendre la recentment desenvolupada funció densitat d'energia cinètica, la qual al comportar-se matemàticament igual a les densitats electròniques regulars, s'ha incorporat en el marc de la semblança. A partir d'aquesta mesura s'han obtingut models QSAR satisfactoris per diferents conjunts moleculars. Dins de l'aspecte del tractament de les matrius de semblança s'ha implementat l'anomenada transformació estocàstica com a alternativa a l'ús de l'índex Carbó. Aquesta transformació de la matriu de semblança permet obtenir una nova matriu no simètrica, la qual pot ser posteriorment tractada per a construir models QSAR.
The present work, even embraced by the Molecular Quantum Similarity Measures (MQSM) theory, is divided into three clearly defined frameworks:
- Creation of Molecular IsoDensity Contours (MIDCOs) from fitted electronic densities.
- Development of an alternative superposition method to replace the maximal similarity rule.
- Quantitative Structure-Activity Relationships (QSAR).
The objective in field of MIDCOs is the application of fitted density functions, initially developed to reduce the computational costs of calculating MQSM, in order to obtain MIDCOs. So, a graphical comparison is carried out using different fittings to different basis sets, derived from calculations carried out at ab initio level. In this way, the visual analogy between the density representations, along with the values of previously calculated similarity measures, which in turn are fully comparable, reinforces the usage of such fitted densities.
Apart from the initial idea, two further studies were done to complement the simple density representations, these are the curvature analysis and the extension to macromolecules. The first step corresponds not only to verify the visual similarity between MIDCOs, but to ensure their coherence in their behaviour at curvature level, allowing to observe inflexion points in the representations and those regions where the density itself presents a convex or concave shape. This first study reveals that both fitted and ab initio densities behave in the same way. In the second study, the method was extended to larger molecules, up to 2500 atoms.
Finally, some of the MEDLA philosophy is applied. Knowing that electronic density rapidly decays as it is measured further from nuclei, its calculation at large distances from the nearest nucleus can be avoided. In this way, a partition of the space was proposed, and the fitted density basis set is only calculated for each atom only in its vicinity. Using this procedure, the calculation time is reduced and the whole process becomes linear with the number of atoms of the studied molecule.
In the chapter devoted to molecular superposition, the creation of an algorithm, along with its practical implementation, called Topo-Geometrical Superposition Algorithm (TGSA), able to direct those alignments that coincide with chemical intuition. The result is an informatics program, codified in Fortran 90, which aligns the molecules by pairs, considering only atomic numbers and coordinates. The complete absence of theoretical parameters allows developing a general superposition method, which provides an intuitive superposition, and it is also relevant, in a fast way without much user-supplied information. Major usage of TGSA has been devoted to provide an alignment to later compute similarity measures, which in turn do not customarily coincide with those values obtained from the maximal similarity rule, most if heavy atoms are present.
Finally, in the last studied subject, devoted to the Quantum Similarity in the QSAR field, three different scopes are treated:
1) Usage of similarity matrices. Here, the so-called similarity matrix, calculated from pairwise similarities in a molecular set, is latterly, and properly manipulated, used a source of molecular descriptors for QSAR studies. With in this framework, several correlation studies have been carried out, involving relevant pharmacological properties, as well as toxicity and physical properties.
2) Application of the electron-electron repulsion energy. This simple contribution consists briefly of taking the value of this magnitude, and by analogy with the self-similarity notation, it is assimilated as a particular case of this measure. This interaction energy is easily obtainable from mecanoquantic software and becomes ideal in order to make a preliminary correlation study, where this magnitude is used as a single descriptor.
3) Calculation of self-similarities, where the density has been modified to account a particular substituent. Previous studies using density fragments, even they provide valuable results, lack of conceptual rigour when isolating a fragment, supposed to be responsible for a particular molecular response, even those densities associated to a particular fragment are already different due to they belong to a common skeleton with different substitutions. A procedure able to fill the gap between usage of density fragments , which avoid the heavy atom effect, and the whole molecular density, which allows to compute a self-similarity, is proposed. It consists of using Fermi Hole density functions, where the holes are defined and averaged around a particular fragment. This procedure medifies the density in a way that is basically collapsed in this molecular bay, but allowing at the same time obtaining a a density function that behaves mathematically in the same way as a regular electronic density; hence it is included in the similarity framework. Those self-similarities computed using this procedure provide good correlations with substituted aromatic acids, allowing to provide an explanation for their acidic behaviour.
From another point of view, conceptual contributions have been developed. A new similarity measure, Kinetic Energy-based, has been implemented, which consists of taking the recently developed Kinetic Energy density function, which has been incorporated into the similarity framework due to it mathematically behaves like a regular density function. Satisfactory QSAR models, derived from this new measure, have been obtained for several molecular systems. Within the similarity matrices field, it has been implemented a new scaling, called stochastic transformation, as an alternative to Carbó Index. This transformation of the matrix allows obtaining a new non-symmetric matrix, which can be later used as a source of descriptors to build QSAR model.

A tuberculose (TB) é uma doença causada pelo Mycobacterium tuberculosis. De acordo com estimativas da Organização Mundial da Saúde, a tuberculose é responsável pela morte de ~2 a 3 milhões de pessoas/ano no mundo e nos próximos 15 anos cerca de 1 bilhão de pessoas deverão ser infectadas, e destas, aproximadamente 35 milhões deverão morrer. Apesar de existirem vários medicamentos sendo utilizados no tratamento da doença, constatasse o crescimento no número de casos devido, principalmente, às variedades resistentes do M. tuberculosis. Considerando-se o aparecimento de cepas resistentes em TB, recomenda-se que novos medicamentos e/ou alvos biológicos alternativos devam ser intensivamente pesquisados. A ribonucleotídeo redutase (RNR), por exemplo, é uma proteína de interesse, pois catalisa uma etapa importante e única na síntese de novo dos dNTPs, reduzindo o ribonucleosídeo 5\' -difosfato ao seu correspondente desoxirribonucleosídeo 5\' -difosfato. A RNR é importante na síntese do DNA, e portanto, na divisão das células. Esta enzima importante, que possuí 16% de homologia com a RNR de mamíferos, é um alvo potencial para o desenvolvimento de novos fármacos, com provável aplicação no tratamento do câncer, da malária e do tripanossoma. Sabe-se que diferentes classes de compostos, através de diferentes mecanismos de ação, inibem a RNR, incluindo as α-(N)-heterocíclicas carboxaldeído tiossemicarbazonas, um dos inibidores mais potentes da RNR. Sabe-se que alguns derivados da tiossemicarbazona, inibidoras da RNR de células tumorais, apresentam atividade frente o M. tuberculosis atuando provavelmente através do mesmo mecanismo, envolvendo a inibição da correspondente RNR. Neste contexto, nesta tese de doutorado, foram aplicadas diferentes abordagens de QSAR/QSAR-3D no estudo de 40 derivados da 2-piridino-carboxaldeído tiossemicarbazona, inibidores da RNR de células H.Ep.-2, retirados de literatura selecionada (French & Blanz-Jr. 1974). Estes compostos foram divididos em cinco séries, a saber: séries A, B, C, D, e E contendo, respectivamente, 40, 39, 30, 23 e 22 compostos, na tentativa de tornar estas séries estruturalmente mais homogêneas. Para cada série, foram criados três grupos de treinamento e os respectivos grupos de teste (I, II e III), visando-se avaliar o poder de predição dos modelos gerados através das análises de QSAR/QSAR-3D. Para as análises de QSAR clássico, foram utilizados como variáveis independentes, os descritores mais relevantes gerados através do programa DRAGON e, pré-selecionados por PLS. Considerando-se a ausência de informações sobre a estrutura cristalográfica da enzima RNR do M. tuberculosis, os estudos de QSAR-3D foram iniciados empregando-se metodologias propostas em CoMFA e, em CoMSIA, implementadas no programa SYBYL. Além destas, foi realizada a modelagem por homologia da RNR do M. tuberculosis, utilizando-se o programa WHATIF. Para as abordagens CoMFA e CoMSIA as geometrias otimizadas através do método semi-empírico AM1 foram alinhadas átomo-a-átomo e, através da similaridade dos respectivos campos estéricos e eletrostáticos, utilizando-se o programa SEAL. Nos dois procedimentos a geometria do composto não substituído, um dos mais ativos na série, foi utilizada como molde considerando-se a ausência de informações sobre a conformação bioativa. A modelagem da RNR por homologia foi realizada utilizando-se como molde as estruturas cristalográficas, respectivamente, do C. ammoniagenes (código PDB 1KGN) e da S. typhimurium (código PDB 1R2F), sambas apresentando valores de identidade superior a 65%. Mais recentemente foram publicados os dados cristalográficos para a cadeia beta (subunidade menor) da RNR do M. tuberculosis (código PDB 1UZR). Os modelos CoMFA e CoMSIA gerados apresentaram valores aceitáveis para os coeficientes de correlação de predição, com altos valores para os coeficientes de correlação ajustados e baixos valores para os erros padrões. Os melhores modelos CoMFA e CoMSIA foram obtidos considerando o grupo com substituintes apenas na posição 5 do anel piridínico. Razoáveis coeficientes de correlação de predição para os modelos CoMSIA com altos coeficientes de correlação de ajuste e baixos valores para os erros padrões forma obtidos. Os mapas de contorno gerados em CoMFA e CoMSIA sugerem que grupos aceptores de ligações de hidrogênio próximos ao nitrogênio do anel piridínico deverá aumentar o valor da atividade inibitória. Esta observação está em boa concordância com os dados da literatura, na qual a formação de um complexo entre a TSC e o íon Ferro foi sugerido para a inibição da RNR. Estes estudos deverão permitir um melhor entendimento sobre as características estruturais desta classe de TSC inibidoras da RNR, como agentes antitumorais, em termos dos campos estéricos, eletrostáticos, hidrofóbico, doador e aceptor de ligações de hidrogênio, bem como a contribuição para o desenvolvimento racional de novos inibidores para esta importante enzima. Adicionalmente, dois compostos preparados em nosso laboratório, demonstraram atividade frente o M. tuberculosis, em testes realizados in vivo.
Tuberculosis is an illness caused by Mycobacterium tuberculosis. Data from World Health Organization (WHO) estimates, that about 2-3 millions of human population died by Mycobacterium tuberculosis infection and that during the next 15 years about 1 billion will be infected and 35 million will certainly die. Although, in the clinic it was found several antiTBdrugs, these numbers will increase due several reasons including M. tuberculosis resistant strains. It has been stressed the importance of novel medicines and/or alternative biological targets research projects. It is known that Ribonucleotide reductase (RNR), is an enzyme that catalyses the rate limiting step in the de novo synthesis of dNTPs, reducing the ribonucleoside 5\'-diphosphates to the corresponding deoxyribonuc1eoside 5\' -diphosphates. RNR has a critical role in the DNA synthesis and, hence, cell division. This key enzyme, that shows 16% homology when compared with mammals RNR, is a potential target for drug design of cell growth inhibitors, with potential application in cancer therapy, antimalaria and trypanosome chemotherapy. It is known that different types of compounds or species by means of different mechanism pathways can show RNR inhibition, including α-(N)-heterocyclic carboxaldehydes thiosemicarbazones that are one of the most potent classes of RNR inhibitors. More than that, some of them, that shows activity against M. tuberculosis seems to follow the same mechanism pathways proposed to the thiosemicarbazones tumor cells activity that means, that they probably are RNR inhibitors. In this study, a series of 40 α-(N)-2-formyl-pyridine thiosemicarbazone derivatives tested against RNR of H.ep.-2-cells (human epidermoid carcinoma), taken from selected literature (French & Blanz-Jr. 1974), has quantitatively analyzed by means of several QSAR/3D-QSAR approaches. These compounds were divided into 5 individual subsets, namely A, B, C, D, and E, having 40, 39, 30, 23 e 22 compounds, respectively. This procedure has been done in order to achieve more structurally homogeneous subsets. For each set, three individual training and test sets (I,II and III) have been created in order to evaluate the predictivity power of the generated QSAR/3D-QSAR models. QSAR analysis have been done using descriptors generated by DRAGON program that have been further pre-selected by PLS procedures. Considering that crystallographic data of RNR M. tuberculosis are not available in the literature, 3D-QSAR studies have been done these applying, initially, CoMFA and CoMSIA approaches, implemented in SYBYL. Homology model studies have been performed with WHATIF program CoMFA e CoMSIA approaches used optimized geometry obtained by semi-empirical AM1 methods that have been aligned by two different methods. Rigid alignment, in which the compounds were fitted atom-by-atom onto a template, based on the root mean square fit. The N(l) and C(2) atoms of the pyridine moiety and the heavy atoms of thiosemicarbazone backbone of TSC were used as template structure. (2) Field based, in which the steric and electrostatic fields, generated by the SEAL program were considered in the alignment. In both procedures the unsubstituted 2-formylpyridine thiosemicarbazone in its syn conformation, has been taken as template. Homology RNR models were done using as template crystallographic data of ammoniagenes (1KGN) and S. typhimurium (1R2F) as template, respectively, with identity larger than 65%. More recent1y new crystallographic data have been published for the beta chain (smaller subunity) of RNR do M. tuberculosis (1UZR). CoMFA and CoMSIA generated models showed acceptable predictive correlation coefficients with high fitted correlation coefficients and low standard errors. Betler CoMFA and CoMSIA models have been derived considering a homogeneous subset of TSC substituted only at 5-position in pyridine ring. Reasonable predictive correlation coefficients for CoMSIA models with high fitted correlation coefficients and very low standard errors were obtained. The derived CoMFA and CoMSIA countour maps suggested that a hydrogen bond acceptor near the nitrogen pyridine ring could enhance inhibitory activity value. This observation is in good agreement with literature, in which a complex formation between TSC and iron ion has been suggested, to RNR inhibition. These studies are expected to enhance the understanding of the structural features of this class of TSC-RNR inhibitors as antitumor agents in terms of steric, electrostatic, hydrophobic and hydrogen donor and acceptor fields as well as to contribute to rational design of inhibitors of this key enzyme. Additionally, two compounds that have been prepared by us showed activity against M. tuberculosis using in vivo test system.

La intensificación de procesos consiste en el desarrollo de equipos y técnicas innovadoras que ofrecen mejoras sustanciales en el proceso, principalmente mediante la disminución del volumen del equipo, consumo de energía o generación de residuos, dando lugar a tecnologías más baratas, seguras y sostenibles. Esta tesis enfatiza la intensificación de procesos como estrategia en la recuperación de dióxido de azufre mediante el empleo de la tecnología de membranas y de líquidos iónicos como absorbente con objeto de eliminar las pérdidas de disolvente.La intensificación del proceso se lleva a cabo en dos etapas:i) Sustitución del equipo convencional (e.g. scrubbers) por un sistema de membranas para eliminar el arrastre de gotas y,ii) Sustitución del disolvente de absorción (N,N-dimetilanilina) por líquidos iónicos para eliminar pérdidas de disolvente por su volatilización en la corriente de gas debido a su presión de vapor despreciable. La selección de un líquido iónico adecuado se basa en su afinidad hacia dióxido de azufre, baja viscosidad, bajo coste y baja ecotoxicidad.En resumen, esta tesis es el primer trabajo que combina el empleo de un contactor de membranas de fibra hueca con líquidos iónicos, contribuyendo al desarrollo de procedimientos innovadores para intensificar el proceso de absorción de dióxido de azufre.
Process intensification consists of the development of innovative devices and techniques that offer significant improvements in chemical manufacturing and processing, decreasing substantially equipment volume, energy consumption, or wastes, and ultimately leading to cheaper, safer and sustainable technologies. This thesis emphasizes the process intensification as the strategy to the recovery of sulfur dioxide, according to the material efficiency and environmental protection, by means of technology based on membranes and ionic liquids as absorption solvents in order to avoid solvent losses.Process intensification is performed in two steps:i) Substitution of conventional equipment (e.g. scrubbers) for a membrane device to avoid drops dragging and,ii) Substitution of the absorption solvent (N,N-dimethylaniline) for ionic liquids to avoid solvent losses due to volatilization of solvent into the gas stream because of their negligible vapor pressure. Selection of a suitable ionic liquid is based on its affinity towards sulfur dioxide, low viscosity, low cost and low ecotoxicity.Thus, this thesis is the first work that combines a hollow fibre membrane contactor and ionic liquids, contributing to the development of innovative procedures to intensify the sulfur dioxide absorption process.

博士
國立臺灣大學
農業化學系
84
25種有機磷殺蟲劑在液相色層分析儀中的滯留時間﹐可以用化合物化學結 構量值計算方法﹐建立一個量化結構及滯留時間關聯式(QSRR)。色層分析 管柱係採用逆相管柱﹐化學結構量化係以分子連接值計算。在甲醇/水 = 90/10 的溶劑相中可獲得較好的量化化學結構及滯留時間關聯式﹐即 Y(90%CH3OH) = －0.3283c + 0.487 0cv－0.154 2cv ﹐其相關係數 (r) 高達 0.991。此結果可將其應用在對其他有機磷殺蟲劑於該移動相的層析 系統中滯留時間的推測。以有機磷殺蟲劑抑制膽鹼酯的情形﹐做為量化結 構及生物活性關聯性研究生物活性指標﹐亦可以用分子連接值做為量化有 機磷殺蟲劑化學結構的方法﹐建立一個可以準確預測有機磷殺蟲劑抑制酵 素情形的數學式﹐本研究獲得一關係式﹐即 Y(pI50) = 0.149 1c ﹐其相 關係數 r為 0.927﹐可預測有機磷殺蟲劑對酵素的抑制。在建立有機磷殺 蟲劑對吳郭魚毒性關聯式時﹐分別以其在層析系統中的滯留時間及對膽鹼 酯酵素抑制程度做為量化化合物的參數﹐可據以建立有機磷殺蟲魚毒性量 化結構關聯式﹐結果獲得一準確性極高的多變數關聯式﹐即 Y(log( LC50)) = 4.306X(I50) +1.295X(90%MeOH) －0.302X(80%MeOH)﹐其相關 係數 r = 0.989。藉由對致毒過程的了解﹐此式中顯現出有機磷殺蟲劑在 層析系統中的滯留時間及其對酵素的抑制情況隱藏著其生物毒性強弱的訊 息。由此關係式顯示若欲獲得有機磷殺蟲劑之魚毒性強弱﹐可省去繁複的 生物檢定試驗﹐而代之以測定其滯留時間及對酵素活性抑制情形來表示。 以分子連接值為量化有機磷殺蟲劑化學結構的方式﹐所建立之有機磷殺蟲 劑魚毒性強弱關聯式﹐同樣的可以有準確的預測結果﹐此關係式為 Y( log(LC50)) = 0.217 0cv ﹐其相關係數 r = 0.976。顯示透過數學方法 計算化合物結構的分子連接值﹐即可準確表示出該結構式之魚毒性的大小 。因此分子連接值應可做為化學物的一種特性值﹐藉由此數值可方便而直 接的透露出隱藏在化學結構中包括毒性等生物活性在內的某些訊息。 A quantitative structure-retention relationships (QSRR) of twenty-fiveorganophosphorus insecticides could be established by the determination ofretention time of organophosphorus insecticide in high performance liquid chromatography and calculation of their chemical structure parameter. Reverse phase column was used for chromatography and structure parameter was quantified with molecular connectivity indices. The best relationship between the retention time and molecular connectivity indices was found by using methanol : water = 9 : 1 as the mobile phase in the HPLC, the quantitative structure- retention relationship with a correlation coefficient (r) of 0.991 are : Y(90%CH3OH) =－0.328 3c＋0.487 0cv－0.154 2cv In the other hand, the biological activity which inhibiting 50% of enzyme activity (I50) were also exhibited a close relationship to the quantitative structure. A quantitative structure- activity relationship (QSAR) with correlation coefficient ( r ) of 0.927 was found as: Y(pI50)=0.149 1c In addition, when examine the relationship between of biological activity, LC50 and I50, for the chemicals with their retention time in HPLC, a multiple linear regression equation with the correlation coefficient of 0.989 are observed as: Y(log(LC50)) = 4.306X( I50) + 1.295X(90%MeOH) - 0.302X(80%MeOH) This relationships imply that some information about the toxicity of organophosphorus insecticides may be hide in the chromatographical retention time and enzyme inhibition. Besides, using molecular connectivity indices to predict the toxicity of organophosphorus to fish were also established in this study. Following equation with correlation coefficient 0.976 are suitable. Y(log(LC50)) = 0.217 0cv Bioactivity of a chemicals seem to be predictable with sturcture molecular connectivity indices.

博士
國立臺灣大學
農業化學系
85
Physical and chemical properties of chemical compounds as well as their bioactivities alter due to the changes in their chemical structures. Therefore, the quantitative structure parameters can be used to describe the physical and chemical properties of different compounds and their bioactivities.In the present study, hydrogen bond acidity index (HAI), hydrogen bond basicity index (HBI), molecular weight (MW) and molecular connectivity indices are selected as quantitative structure parameters. The capacity factors (k'') of 19 carbamate pesticides are determined on the reversed-phase column with mobile phases of methanol/water or acetonitrile/water in various ratios. By the use of multiple regression analysis, the ln k'' values are screened against the quantitative structure parameters for all possible two-, three- and four-variable combinations. With different mobile phase systems, the best relationships are obtained to the 4-variable, the multiple correlation coefficients are between 0.898-0.934. The independent variable HBI in all regression equations means that HBI is an effective parameter in retention of carbamate pesticides in reversed-phase high-performance liquid chromatographic(RP-HPLC).The inhibition of acetylcholinesterase by carbamate pesticides can be explained with affinity constant (Ka) and carbamylation constant (kc), in the other word, in explanation of inhibition constant (Ki). The natural logarithm value of these three constants can be described by the quantitative structure parameters. The best relationships are obtained to the 4-variable, the multiple correlation coefficients are between 0.911-0.946. Through the further study, (1／0XN)2 and (R1 3XcV) are effective parameters to explain the inhibition of carbamate pesticides on acetylcholinesterase. The natural logarithm values of median lethal concentration (LC50) of carbamate pesticides to Daphnia pulex can also be described by the quantitative structure parameters, the best relationship is obtained to the 4-variable, the multiple correlation coefficient is 0.888.The above mentioned regression equations are tested by random effect and F-test. The regression coefficients are also tested by t-test. All of the statistic results are significant. The values of these quantitative structure parameters are sure to reveal certain relationship among ln k'', ln Ka, ln Ki, ln kc and ln (LC50).

Chemical hardness (eta), calculated by density functional theory (DFT), was firstly used as one of the chemical reactivity descriptors to set up the one descriptor 2D-QSAR model of platinum drugs. In this simple but promising model, the antitumour activities (log GI50) evaluated by National Cancer Institute (NCI) of structure-based groups containing normal sp 3 nitrogen and R,R-diamminecyclohexane (R,R-DACH) as the ligand showed good correlation. It was also demonstrated that silane and stereoisomers of DACH groups showed special patterns. This study also made use of the COMPARE program from NCI to evaluate the activity profile and the analysis of the data revealed these distinct patterns are influenced by the mechanism of the drugs.
Computer-aided drug design (CADD) techniques have been applied to establish quantitative structure-activity relationships (QSAR) and quantitative structure- property relationships (QSPR) models. Although these techniques are widely used in organic drugs, new metal-based drugs were hindered from development for lack of metal parameters, such as potent new platinum drugs as a major group of drugs used in cancer treatment. The purpose of the present study, therefore, is to generate novel platinum parameters based on previous work and then set up the simple QSAR/QSPR model with predictive abilities.
Finally, two 3D-QSAR and 3D-QSPR models obtained using Sybyl software. One was for demethylcantharidin (DMC) analogues as phosphatase 2A (PP2A) inhibitors. The other was describing the hydrophobicity of platinum drugs. In this research, the platinum atom was introduced to Sybyl and thus made it possible for the first time to use comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) methods to investigate platinum drugs. All 3D models indicated good predictive ability and thus provided an effective method to design new potent platinum drugs.
To clarify the pattern of stereoisomers of the DACH group, new platinum parameters was introduced to the AMBER software successfully. Moreover, stereoisomers of the DACH group which formed 1,2-GG intrastrand cross-links with DNA were studied by molecular dynamics (MD) simulations using AMBER. The calculated binding energies between R,R-DACH-Pt, S,S-DACH-Pt and cis-DACHPt moieties and DNA revealed a strong correlation with antitumour activities. The result provided more clues to understand the biological interactions of chiral platinum drugs. DNA structure analysis indicated that DNA tolerated the distortion resulted in the different Pt-DNA adducts and various local and global structure distortions were found. Natural bond orbital (NBO) analysis of hydrogen bonding on Pt-DNA adducts at a AGGC site revealed that R,R-DACH-Pt moiety alleviated the repulsion by unwinding the DNA, whereas the S,S-DACH-Pt adduct avoided the interaction by distorting the H bonds of binding site basepairs. Hence, the structural differences of chiral platinum drug led to its distinct activity.
Yang, Lifeng.
"June 2008."
Advisers: Steve C. F. Au Yeung; Yee-Ping Ho.
Source: Dissertation Abstracts International, Volume: 70-03, Section: B, page: 1541.
Thesis (Ph.D.)--Chinese University of Hong Kong, 2008.
Includes bibliographical references (p. 159-172).
Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web.
Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web.
Abstracts in English and Chinese.
School code: 1307.

Leung Chung Wai.
Thesis (M.Phil.)--Chinese University of Hong Kong, 2006.
Includes bibliographical references (leaves 142-146).
Abstracts in English and Chinese.
ABSTRACT (ENGISH) --- p.iii
ABSTRACT (CHINESS) --- p.v
ACHKNOWLEDGEMENTS --- p.vii
TABLE OF CONTENTS --- p.viii
Chapter CHAPTER 1 --- Introduction and Background
Chapter 1.1 --- Introduction of Platinum Drugs --- p.1
Chapter 1.2 --- Mechanism of Action of Cisplatin --- p.3
Chapter 1.3 --- Structure-Activity Relationships of the Platinum Drug 、 --- p.4
Chapter 1.4 --- QS AR Parameters --- p.9
Chapter 1.4.1 --- Chemical Hardness: Descriptor of Chemical Reactivity --- p.9
Chapter 1.4.2 --- Possible Reaction Pathway of Platinum Drugs --- p.12
Chapter 1.4.2.1 --- Proposed DNA Binding Pathway of Platinum Drugs --- p.13
Chapter 1.4.2.1.1 --- Hydrolysis Pathway --- p.13
Chapter 1.4.2.1.2 --- DNA Binding Pathway Involving the S-containing Biomolecules (Methionine Pathways) --- p.16
Chapter 1.4.2.1.3 --- Conclusion --- p.21
Chapter 1.5 --- Thesis Scope --- p.22
Chapter CHAPTER 2 --- Theory and Methodology
Chapter 2.1 --- Introduction --- p.24
Chapter 2.2 --- Density Functional Theory (DFT) --- p.24
Chapter 2.2.1 --- Kohn-Sham Theorem --- p.25
Chapter 2.2.2 --- Exchange-Correlation Energy Functional --- p.27
Chapter 2.3 --- Basis Set --- p.27
Chapter 2.3.1 --- Relativistic Effective Core Potential --- p.27
Chapter 2.3.2 --- Double-Zeta --- p.28
Chapter 2.3.3 --- Polarized Basis Set --- p.29
Chapter 2.4 --- Solvation Model --- p.30
Chapter 2.4.1 --- Continuum Model --- p.30
Chapter 2.4.1.1 --- Simple Solvation Model --- p.31
Chapter 2.4.1.1.1 --- Electrostatic Component --- p.31
Chapter 2.4.1.1.2 --- Dispersion-Repulsion Interaction --- p.33
Chapter 2.4.1.1.3 --- Cavitatoin Energy --- p.35
Chapter 2.4.1.2 --- Polarized Continuum Model --- p.36
Chapter 2.5 --- Methodology --- p.39
Chapter 2.5.1 --- Calculation of DFT Global Reactivity Index --- p.39
Chapter 2.5.1.1 --- Calculation for the Reaction Intermediates --- p.41
Chapter 2.5.2 --- Calculation of the Reaction Pathways --- p.42
Chapter CHAPTER 3 --- Results and Discussion
Chapter 3.1 --- Introduction --- p.49
Chapter 3.2 --- Optimized Structure against Experimental Geometry --- p.49
Chapter 3.3 --- Kohn-Sham Orbitals --- p.54
Chapter 3.3.1 --- Location of the HOMO and LUMO --- p.55
Chapter 3.4 --- Results of the DFT Reactivity Parameter --- p.57
Chapter 3.5 --- Chemical Structure of the Drugs in the QSAR --- p.64
Chapter 3.6 --- QSAR Analysis --- p.67
Chapter 3.6.1 --- The Overall QSAR Plot of the Platinum Drugs --- p.68
Chapter 3.6.1.1 --- Empirical Applicability of the QSAR on the Platinum(IV) Drugs --- p.70
Chapter 3.6.1.2 --- Detail QASR Study According to the Type of Platinum Drug --- p.71
Chapter 3.6.1.2.1 --- QSAR Study of the non-“trans-DACH´ح Platinum Drugs --- p.72
Chapter 3.6.1.2.1.1 --- "QSAR Equation of the non-""trαns-DACH"" Platinum Drugs" --- p.75
Chapter 3.6.1.2.2 --- QSAR Analysis for the Pt-trαns-DACH Drugs --- p.77
Chapter 3.6.1.2.2.1 --- "QSAR Study of trans-S,S-DACH Platinum Drugs" --- p.79
Chapter 3.6.1.2.2.2 --- "QSAR Study of trans-R,R-DACH Platinum Drugs" --- p.80
Chapter 3.6.1.3 --- Summary --- p.81
Chapter 3.7 --- QSAR Study of the Important Intermediates Using Chemical Hardness --- p.82
Chapter 3.7.1 --- Optimized Structure for the Intermediates --- p.84
Chapter 3.7.2 --- QSAR of the Dichloride Pt-Drugs Using Chemical Hardness of Parent Compounds --- p.90
Chapter 3.7.3 --- QSAR of the Dichloride Pt-Drugs Using Chemical Hardness of Hydrolysis Intermediates --- p.91
Chapter 3.7.4 --- QSAR of the Dichloride Pt-Drugs Using Chemical Hardness of Cyclic-Methionine Intermediates --- p.93
Chapter 3.7.5 --- Conclusion --- p.95
Chapter CHAPTER 4 --- Results and Discussion
Chapter 4.1 --- Introduction --- p.96
Chapter 4.2 --- Study Scheme --- p.97
Chapter 4.3 --- Optimized Structures --- p.98
Chapter 4.4 --- Comments on the Reliability of the Calculation Model --- p.103
Chapter 4.4.1 --- Reaction Profile in the Gas Phase --- p.104
Chapter 4.4.2 --- Reaction Profiles Using Simple Solvation Model --- p.105
Chapter 4.4.2.1 --- Defects of the Simple Solvation Model --- p.107
Chapter 4.4.3 --- Reaction Profile Using PCM-UAHF Solvation Model --- p.109
Chapter 4.4.3.1 --- Selection of the Reaction Parameters for the QSAR Study --- p.112
Chapter 4.5 --- QSAR Study of Platinum Drugs Using the Reaction Parameters (AG and ΔG+) --- p.121
Chapter 4.5.1 --- QSAR Analysis Using ΔG+(hydrolysis) --- p.121
Chapter 4.5.2 --- QSAR Analysis Using ΔG(hydrolysis) --- p.123
Chapter 4.5.3 --- QSAR Analysis Using ΔG+(guanine) --- p.125
Chapter 4.5.4 --- QSAR Analysis Using ΔG(guanine) --- p.127
Chapter 4.5.5 --- Further investigation of the Bidentate Pt-drugs DNA Binding --- p.129
Chapter 4.5.5.1 --- Calculation Model --- p.129
Chapter 4.5.5.2 --- Bidentate Pt-Drugs Reactions --- p.130
Chapter 4.5.5.3 --- Selection of the Calculated Model for the QSAR Study --- p.133
Chapter 4.5.5.4 --- QSAR Analysis Using ΔG+(guanine) for the Platinum Drugs with Bidentate Caboxylate Ligands --- p.136
Chapter 4.5.5.5 --- QSAR Analysis Using ΔG(guanine) for the Platinum Drugs with Bidentate Carboxylate Ligands --- p.137
Chapter 4.5.6 --- Conclusion --- p.138
Chapter CHAPTER 5 --- Conclusion Remarks and Future Works
Chapter 5.1 --- Conclusion --- p.140
Chapter 5.2 --- Future Works --- p.141
REFERENCES --- p.142

碩士
國立中興大學
土壤環境科學系所
97
The environment is often exposed to chemical mixtures from multiple sources. The toxicity of various chemical mixtures is higher than single chemicals. However, the vast majority of toxicity studies deal with single chemicals, and therefore the prediction of mixture toxicity becomes a necessary and vital issue. In recent years, the development of quantum mechanical theory was combined with the progress of computational technology, which means quantitative calculation can be conducted from the atomic or molecular structure of a substance with little or even without empirical results. Besides, the parameter calculated was directly connected to organic activity, toxicity, chemical reaction, to construct the projection patterns, among which the QSAR is generally used to project mixture toxicity now. In this study, the objective is the binary mixtures toxicity of 12 benzene and its derivatives in the environment and 9 organophosphorus pesticides used with high frequency domestically, from which the DFT of quantum mechanical theory developed in recent years is used as a basis to build up the QSAR of toxicity prediction. The differences between Semi-empirical (AM1) and DFT (B3LYP) are discussed as well, and the prediction pattern will further be applied to each field to access mixtures and reach the goal of fast prediction. The results suggest that the results of prediction pattern are similar from either B3LYP or AM1 that are used to calculate benzene and its derivatives and mixtures. When using one parameter to predict toxicity, total surface area (TSA), apolar surface area (APSA), electron affinity (EA), and chemical potential (μ) are major factors. As multi-parameters are concerned, the increase of reaction energy (ΔEAB) and global soft (S) are required as parameters. Thus, AM1 has priority for choices in the future due to its fast calculation. When constructing prediction patterns of toxicity, surface area is an important parameter no matter benzene and its derivatives or organophosphorus pesticides are concerned with single parameter. And surface area could be influenced by chemical polarity depending on different subjects of prediction. Moreover, groups of fat and fragrance are factors to influence toxicity as well, so the number of ring and atom are used as parameters while the variety of toxic substances are complicated. When predicting mixture toxicity with multi-parameters, TSA, Etot, η, S, and μ are necessary parameters. And avoid aromatic compounds induce difference, so use sum of ring (R) and sum of atoms (NO, NN, NS, NP, NCl). Besides, ΔEAB is an important parameter in mixture.

Hepatotoxicity is a serious adverse health effect caused by drugs and other chemical toxins generally detected in the later stages of drug development or in whole animal studies. Thus, development of screening approaches available for earlier identification of hepatotoxic molecules is necessary. A novel in vitro- in silico test system for the evaluation of the molecular mechanisms of xenobiotic toxicity in primary hepatocyte systems is presented here. It is well established that hepatocytes in vitro are most representative of hepatotoxicity in vivo, and are most useful for the determination of xenobiotic hepatotoxicity mechanisms at the molecular and cellular level. There is an on-going interest in Quantitative Structure-Activity Relationships (QSAR) in toxicology, as it can identify correlations between chemical structure and biological activity. QSAR can be used to evaluate the effects of metabolism and toxicity as many physicochemical descriptors reflect simple molecular properties that can provide insight into the physicochemical nature of the activity under consideration. QSARs were determined for hepatotoxicity of halobenzenes, p-benzoquinones, α,β-unsaturated carbonyl compounds and nitroaromatics towards isolated hepatocytes. A molecular link was established for their proposed toxicity pathways. For example oxidative activation was linked to EHOMO (energy of the highest occupied molecular orbital) values and hydrophobicity (log P) of the chemicals, while reductive activation was linked with ELUMO (energy of the lowest molecular orbital) values and log P. Such relationships may thus be useful for predicting toxicity of other chemicals of the same mechanism of toxicity. Due to the complexity involved in the phenomena of hepatotoxicity, unravelling of structure-hepatotoxicity relationships is a complicated task. A conceptual framework for QSAR modeling is proposed that involves recognition of molecular initiating events as potential endpoints to improve the prediction potential of QSAR models. Acute toxicity of reactive chemicals could be based on an initial reaction with biomolecules, thus the theory of covalent binding reactivity was used to test this concept. Reactivity assays with thiol and amine surrogate nucleophiles were used to determine susceptibility to toxicity. The derived QSAR expressions suggested that covalent binding reactivity is a good correlate to hepatotoxicity, however only if electrophilicity was the main mechanism of toxicity.

碩士
國立成功大學
化學系碩博士班
91
The acute toxicity to Pseudomonas putida (ATCC 23973) of 66 chlorinated compounds, containing phenols, benzenes, anilines and toluenes, was using log(1/Ki) as a toxicity parameter. The Ki value was estimated from Pseudomonas initial oxygen uptake (PIOU) method. Except for five monosubstituted benzenes, all of the testing compounds are noncompetitive inhibitors. Quantitative Structure-Activity Relationships (QSARs) were developed for chemical and toxicological subsets. A quantitative structure-toxicity model correlating with 1-octanol/water partition coefficient (log P) and energy of the lowest unoccupied molecular orbital (LUMO) was developed and result to the following equation. log(1/Ki)=-0.38(0.04)LUMO+0.46(0.03)logP-2.19(0.08) n = 61, R2 = 0.896, s = 0.176, F = 250.0 Two chlorophenols (pKa≦6.5) being classified as “uncouplers” were outliers of our QSAR. The logP of these phenols was rectified by the corresponding phenoxide. The model was changed to log(1/Ki)=-0.37(0.04)LUMO+0.48(0.02)logP-2.25(0.07) n = 61, R2 = 0.923, s = 0.152, F = 347.32 The lower toxicity of these outliers is due to the dissociation of phenols to their corresponding phenoxides rather than uncoupling mechanism of action on the synthesis of ATP and oxidative phosphorylation. The current results also indicated that phenols, benzenes, anilines and toluenes are share a similar inhibitory action mechanism as they can be explained by the same QSAR equation. This result is different from that of Tetrahymena pyriforms and Microtox.

碩士
中興大學
土壤環境科學系所
99
QSARs are currently being applied in many disciplines, with many pertaining to the study of toxicological activities of environmentally important molecules and drug design. Nevertheless, there is still room to improve the QSARs to become a tool for studying the mechanism. But there are some problems in QSAR that it can’t become a tool for studying the mechanism. Therefore, this study expected QSAR to become a research mechanism tools through descriptors of electronic attributes established integral information of chemical reactions. Electronic has four different attributes (contact, non-contact, deformation, non-deformation), pairwise combine to four major chemical bonding forces (covalent bond, ionic bond, where der Waals force and hydrophilic interactions). The aim of the present study was to represent an attempt to correlate the electronic attribute-based descriptors with the experimentally determine the toxicity of substituted Organic Compounds to the Photobacterium phosphoreum. The results showed that the main toxicity mechanism of benzene and naphthalene compounds to Photobacterium phosphoreum is ionic bond and covalent bond , respectively. In addition, the QSAR models are predictive and descriptors of electronic attributes establish chemical reaction descriptors (chemical bonding) integrity. In summary, it can be concluded from the information presented in this study that a model based on electronic attributes not only can calculate the toxicity of any hypothetical compound of the series, but also can identify mechanism of organic compounds.

碩士
國立交通大學
環境工程系所
102
This study present the toxicity data of ten organothiophosphorus compounds to Daphnia magna using 48-hours acute toxicity test. Results indicate that Phoxim is the most toxic compound, the reason caused higher toxicity of Phoxim is dependent on its higher charge density due to benzyl cyanide group. Quantitative structure-activity relationships (QSARs) were established between the EC50 values and various molecular descriptors, and a highly predictive two-variable QSAR models were obtained. According to QSAR models, the toxic mechanism was highly related to hydrophobicity of these compounds, and this result indicated that organothiophosphorus compounds caused toxicity to Daphnia magna mainly due to permeate through membrances, changing the components of the cell membrances. Adding reactive parameters slightly improve the r2 values indicated that the reaction rate of the organothiophosphate with AChE, the conversion of organothiophosphate to organophosphate and the molecule polarity are not rate-limiting steps. The preliminary environmental risk assessment was conducted following the European Union RQ model. Results point out that most of these compounds exist potential risks to aquatic environment. Risk quotient derived from acute and estimated chronic toxicity data were different, using acute toxicity data may overestimate the risk quotient value.

碩士
國立高雄第一科技大學
環境與安全衛生工程系碩士班
106
Many flammable materials are used in industrial processes, so avoiding fire hazards is a critical issue for such processes. Liquid organic compounds are most commonly used chemical during process, we use flash point (FP) to determine what category of flammable liquid they are and decide the storage method through its category. Besides we also understand the combustion characteristics of the chemical through the auto-ignition temperature (AIT), and determine what correct specifications of explosion-proof electric equipment is provided for handling such flammable materials in order to effectively prevent the occurrence of fire and explosion. However, using experimental methods to evaluate physical properties of the chemicals is usually time consuming, costly, and it is even precarious for toxic or explosive chemicals. Therefore, REACH regulation encourages the use of cost effective alternatives, such as the Quantitative Structure Activity Relationships (QSAR) approach. Our study retrieved all data from DIPPR 801 database maintained by the American Institute of Chemical Engineers (AIChE), and selected only experimental data to avoid using predicted value to affect the model results. Eventually, 786 experimental data were used to establish a flash point prediction model with four descriptors. This model gives the goodness-of-fit performance (

碩士
國立成功大學
化學系碩博士班
91
We have used the Pseudomonas putida (ATCC 23973) initial oxygen uptake (PIOU) data to assess the toxicity of the monosubstituted aromatic compounds and phenol derivatives. The inhibition constant, Ki, of monosubstituted aromatic compounds and phenol derivatives were measured and found that most phenol derivatives are belonging to the non-competitive inhibitors toward Pseudomonas putida. These Ki constant represent the concentration of the phenol derivatives, under which 50% of the oxygen uptake activity of Pseudomonas putida has been inhibited. Thus, Ki is used as an index to assess the toxicity. The penetration ability of the toxic chemical into the organism and its following reaction site with the organism are use to assess the toxicity. Thus, the Ki value was correlated with both the n-octanol/water partition coefficient (log p) and the energy of lowest unoccupied molecular orbital (LUMO) to establish a quantitative structure-activity relationship (QSAR) model and found that this established equation allow one to predict the inhibition mechanism toward P.putida. Several phenol derivatives ( pKa≤6.5 ) with lower toxicity previously called uncupler was also found not following the QSAR behavior, i.e. can not be explained using the established QSAR equation. The abnormal toxicity behavior of these compounds was found mainly due to having weak acid dissociation property into its conjugate base rather than due to being an uncoupler inhibiting the oxidative phosphorylation of ADP to form ATP. Comparing our results obtained from PIOU method with that from other test methods, it was found that our results have a good relationship with that from a 2-day Tetrahymena pyriformis with r = 0.92.

碩士
國立交通大學
環境工程系所
95
The objective of this study is to study the toxic effect of benzoic acids on Pseudokirchneriella subcapitata using a closed system test. The effects of benzoic acids were evaluated by three kinds of response endpoints, i.e., cell density, algal growth rate, and the dissolved oxygen production. Median effective concentratons (EC50s) were estimated using the Probit model with a test duration of 48hr. The quantitative structure-activity relationships (QSARs) were established based on the 1-octanol/water partition coefficient (logKow) and the numbers of benzene substituted with hydroxyl revealed a good relationship (R2=0.915~0.965). Benzoic acids with halogens at the meta- and para- positions were more toxic than benzoic acids with halogens at the ortho- position. As for the hydroxylated benzoic acids considered here, the results indicate that ortho- hydroxylated benzoic acids have lower pKa values, but the resluts showed higher toxicities than the meta- and/or para- hydroxylated ones, indicating that the hydroxybenzoate derivatives behaved differently from the halobenzoates. Especially, Tri-hydroxybenzoates may have higher affinity to algae than expected from their low logKow and pKa values. The results also reveal that the value of the lower effect concentration of the benzoic acids (cell density, algal growth rate, and the dissolved oxygen production). This demonstrates that the relative sensitivity is NOEC >EC10>. Besides, the experiment results (EC50) are compared with literature data derived by various toxicity tests. The order of the relative sensitivity is then obtained as follows : Microtox> algae(Final yield)> algae(DO production)> algae(Grwoth rate)> Daphnia magna> Tetrahymena pyriformis.

碩士
國立交通大學
環境工程系所
96
The objective of this study is to study the toxic effect of primary propargylic alcohols on Pseudokirchneriella subcapitata using a closed system test. The effects of primary propargylic acids were evaluated by three kinds of response endpoints, i.e., dissolved oxygen production, cell density and algal growth rate. Median effective concentratons (EC50s) were estimated using the Probit model with a test duration of 48hr. Primary propargylic alcohols with alkene on the second position exhibit higher toxicity than alkene on the third position. The quantitative structure-activity relationships (QSARs) were established based on the 1-octanol/water partition coefficient (logKow) and using log(1/EC50) , 2-propyn-1-ol was not included , revealed a good regression relationship (R2=0.9, based on final yield). The results also reveal that the value of the lower effect concentration of the primary propargylic alcohols (cell density, algal growth rate, and the dissolved oxygen production). This demonstrates that the NOEC value can offer a better standard for protecting our aquatic environment than the EC10 value . Hence, comparing with other aquatic organismsthe acute toxicity data , The order of the relative sensitivity is then obtained as follows : Fathead minnows> P. subcapitata > Tetrahymena pyriformis.

碩士
國立陽明大學
生物化學研究所
89
Abstract Dr. Sheau-Farn Yeh’s lab found that taiwanin A, isolated from Taiwania Cryptomeriodes, is cytotoxic for several cancer cell lines, especially for human hepatoma Hep G2/A2 and Hep 3B/T2 cells. Besides cytotoxicity, the compound can arrest the cells at G2/M phase. Not only taiwanin A and its derivatives, but also compounds which isolated from the plane Clausena excavata and the roots of Rubia cordifolia have been found to have biological activities. The compounds that isolated from Clausena excavata and Rubia cordifolia can reduce the secretion of hepatitis B surface antigen (HBsAg). After being informed of the biological activity of three groups of derivatives, to derive the pharmacophore through the computer to modify original compounds and search new effective compounds form database are substantial problems that we are interested in. In this thesis, we use several descriptors, such as traditional octanol/water partition coefficient, the energies and electronic densities of orbitals calculated by quantum chemistry method, to investigate the relation between the structures and the biological activity of compounds. The results indicate that there is an important relationship between the substructures of butyrolactone in taiwanin A and its biological activity, we find that the enlargement of structure space at the place of methylene of piperonyl group of taiwanin A may be a way to improve its biological activity. We also use the package Catalyst™ through establishing the hypothesis of pharamocophore of these groups of compound to search other effective compounds. Finally, the Monte Carole method is performed to simulate the effect of cell cycle under the treatment of drug. The result predicted by our program shows that the percentage of each phase in cell cycle is quite consistent with that of experimental data. We hope that the present protocol can suggest a new choice of biological activity for describing drug functions.

碩士
國立交通大學
環境工程系所
96
A closed-system algal toxicity test was applied to evaluate the toxicity of secondary and tertiary propargylic alcohols to Pseudokirchneriella subcapitata. The effects of propargylic alcohols were evaluated by three kinds of response endpoints, i.e., dissolved oxygen production, cell density, algal growth rate. Median effective concentraton (EC50) were estimated using the Probit model. The quantitative structure-activity relationships (QSARs) were established based on the 1-octanol/water partition coefficient (logP) and an electronic parameters-lowest unoccupied molecular orbit (Elumo). The result shows that the highest sensitivity is cell density between three end points, and the toxicity of secondary and tertiary propargylic alcohols is correlation to the position of alkynyl and hydrophobicity, respectively. Compare with sensitivity for other species, BOD-FY > BOD-DO > Fathead minnow > BOD-GR > Tetrahymena Pyriformis。 The QSAR about the secondary propargylic alcohols has a good fit (R2 = 0.852~0.926) with the logP and Elumo except three outliers that alkynyl attached to 1-position. The QSAR using logP for tertiary propargylic alcohols has a good fit (R2 = 0.914~0.965). Moreover, a cut-off value approach is proposed to determine whether NOEC or EC10 should be chosen for estimating low toxic effects. The results indicate that NOEC offers better protection to test organisms than EC10.