Orientador: Marcia Miguel Castro Ferreira / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Quimica / Made available in DSpace on 2018-08-15T02:36:54Z (GMT). No. of bitstreams: 1 Melo_EduardoBorgesde_D.pdf: 5803988 bytes, checksum: 6377ba7d02209df67db56420a1568312 (MD5) Previous issue date: 2009 / Resumo: Apesar da implantação da HAART, existe uma necessidade contínua de novos agentes anti- HIV. Os inibidores da enzima HIV-1 integrase (HIV-IN) constituem um dos mais recentes avanços na luta conta a AIDS. A principal abordagem utilizada nessas pesquisas são os métodos relacionados ao planejamento de fármacos auxiliados por computador (CADD). Neste trabalho, foram realizados três estudos QSAR (2D, 4D e híbrido) utilizando um conjunto de treinamento formado por 85 compostos descritos como inibidores da reação de transferência de fita catalisada pela HIV-IN. No estudo QSAR-2D, foram utilizados 1291 descritores físico-químicos obtidos por diversos programas. Para o estudo QSAR-4D, perfis de amostragem conformacionais (PACs) foram obtidos com o programa de dinâmica molecular Gromacs 4, e 65.856 descritores de campo (Coulomb e Lennard-Jones) foram obtidos a partir do programa LQTA-QSAR. As seleções de variáveis foram realizadas pela metodologia Ordered Predictors Selection (OPS), e os modelos foram construídos utilizando regressão por quadrados mínimos parciais (PLS). Na etapa de QSAR-2D, foi realizado um estudo preliminar com 33 compostos com baixa variabilidade estrutural e 167 descritores de mais simples interpretação. O modelo obtido foi formado por duas variáveis latentes e quatro descritores. Esse modelo apresentou uma relação direta com o mecanismo de inibição mais aceito. Já para o modelo com o conjunto completo, foram selecionados quatro descritores, porém de difícil interpretação, provavelmente devido à grande variabilidade estrutural do conjunto de treinamento. Já para o modelo QSAR-4D, uma relação direta com o mecanismo de inibição, com descritores correspondentes à interação com os co-fatores metálicos e com a alça hidrofóbica do sítio de ligação da HIV-IN, também pôde ser traçada. Todos os modelos apresentaram qualidade estatística aceitável, com boas capacidades de predição interna e robustez, além de não apresentarem correlação ao acaso. Já o modelo híbrido, construído com alguns dos descritores selecionados nos estudos anteriores, possui alta qualidade estatística, mas é inferior ao modelo QSAR-4D. Logo, ao serem considerados os resultados obtidos, conclui-se que os objetivos da tese foram alcançados e que os modelos obtidos apresentaram grande potencial para proposição de novos inibidores da HIV-IN. / Abstract: Despite the HAART implantation, there is a continuous need to search for new anti-HIV agents. The HIV-1 integrase (HIV-IN) inhibitors are one of the most recent breakthrough in AIDS research. So, the computer aides-drug design (CADD) related methods have been the main approach used in the research of such class of drugs. In this work, three QSAR studies (2D, 4D and hybrid), with a training set, consisted of 85 inhibitors of strand transfer (ST) reaction catalyzed by HIV-IN. In the 2D-QSAR study, 1,291 physicochemical descriptors were obtained by several programs. For the 4D-QSAR study, the conformational essembles profiles (CEPs) were obtained by the molecular dynamic program Gromacs 4. With the LQTA-QSAR program, 65,856 descriptors (Coulomb and Lennard-Jones) were obtained. In both the studies, the variable selections were carried out according to the Ordered Predictors Selection (OPS) method while the models were composed with Partial Least Squares (PLS) regression. In the 2D-QSAR step, a preliminary study with 33 compounds with low structural variability and 167 descriptors of more simple interpretation was developed. The obtained model was based on two latent variables and four descriptors. But, for the model with a complete set, there were four selected descriptors, although the difficult interpretation, probably due to the great structural variability of the training set. On the other hand, a direct relation with the inhibition mechanism could be traced for the 4D-QSAR model, including descriptors related with the interaction with the metallic co-factors and with the hydrophobic loop, placed in the binding site of HIV-IN. All the models showed an acceptable statistic quality, with good capacity of internal prediction and robustness. Moreover, the models did not present any randomized correlation. But, the hybrid model, built with some of descriptors selected in both studies, although it also has high statistic quality, is inferior to the 4D-QSAR model. Hence, considering the good obtained results, it can be concluded that the purposes of this thesis were achieved and that the models present a great potential to propose new HIV-IN inhibitors. / Doutorado / Físico-Química / Doutor em Ciências

Quaternary ammonium salts are surface active bactericides. A mechanism of their biological activity has been well studied experimentally, and it encompasses two stages. The first stage involves electrostatic interactions of polar functional groups of the salts with oppositely charged functional groups on a bacterial cell surface, and the second stage includes incorporation of their lipophilic groups into a bacterial cell membrane. However, despite numerous experimental studies, computational modelling of this mechanism with the aim to support experimental observations with theoretical conclusions, to the author's knowledge, has not yet been reported. In the current study, linear regression models correlating theoretical descriptors of lipophilicity and electronic properties of mono- and disubstituted imidazolium carboxylates with their biological activity towards Escherichia coli have been developed. These models established that biological activity of these salts is governed by the chemical structures of imidazolium cations, and that the centre of this biological activity is located in the long alkyl side chains of the cations. It was also found that these side chains have an intrinsic electronic potential to form internal C-H- -H-C electrostatic interactions when their lengths reach seven carbon atoms. Additionally, the nature of the C-H- -O-C inter-ionic electrostatic interactions in imidazolium carboxylates has been explored via a topological analysis of these interactions in 1-ethyl-3-methylimidazolium acetate. Thus, it was established that these electrostatic interactions are hydrogen bonds.

Background: In drug discovery morphological profiles can be used to identify and establish a drug's biological activity or mechanism of action. Quantitative structure-activity relationship (QSAR) is an approach that uses the chemical structures to predict properties e.g., biological activity. Support Vector Machine (SVM) is a machine learning algorithm that can be used for classification. Confidence measures as conformal predictions can be implemented on top of machine learning algorithms. There are several methods that can be applied to improve a model’s predictive performance. Aim: The aim in this project is to evaluate if ligand-based modelling can be improved by combining features from chemical structures, target predictions and morphological profiles. Method: The project was divided into three experiments. In experiment 1 five bioassay datasets were used. In experiment 2 and 3 a cell painting dataset was used that contained morphological profiles from three different classes of kinase inhibitors, and the classes were used as endpoints. Support vector machine, liblinear models were built in all three experiments. A significant level of 0.2 was set to calculate the efficiency. The mean observed fuzziness and efficiency were used as measurements to evaluate the model performance. Results: Similar trends were observed for all datasets in experiment 1. Signatures+CDK13+TP which is the most complex model obtained the lowest mean observed fuzziness in four out of five times. With a confidence level of 0.8, TP+Signatures obtained the highest efficiency. Signatures+Morphological Profiles+TP obtained the lowest mean observed fuzziness in experiment 2 and 3. Signatures obtained the highest correct single label predictions with a confidence of 80%. Discussion: Less correct single label predictions were observed for the active class in comparison to the inactive class. This could have been due to them being harder to predict. The morphological profiles did not contribute with an improvement to the models predictive performance compared to Signatures. This could be due to the lack of information obtained from the dataset. Conclusion: A combination of features from chemical structures and target predictions improved ligand-based modelling compared to models only built on one of the features. The combination of features from chemical structures and morphological profiles did not improve the ligand-based models, compared to the model only built on chemical structures. By adding features from target predictions to a model built with features from chemical structures and morphological profiles a decrease in mean observed fuzziness was obtained.

The cost of developing new drugs is high and the aim of computer-assisted drug discovery is to reduce that development cost, either through virtual screening or generating novel compounds. System biology is one approach to drug discovery where the response of a biological system is the subject of study, instead of drug target interaction. One way to observe a biological system is through microscopy images that are taken of cells perturbed with compounds. Image software extracts information called morphological profiles from the images that can be used for data hungry models. One of the ways artificial intelligence has been applied to drug discovery is with generative models that can generate new compounds. One such generative model is reinforcement learning that employs a critic to guide the generation of compounds towards desirable behaviors. In this study different machine learning models were tested if they could predict the morphological response of COVID-19 infected cells to compounds from their structure. No modells showed any promising results. The reason that no model performed well was because of the dataset. There is a lot of variance in the dataset, meaning that the response to the same compound varies. There was also a lot of difference between the compounds in the dataset, meaning that any representation that the model learns does not transfer over to other compounds. The data set was also imbalanced with more inactive compounds.

京都大学 / 新制・課程博士 / 博士(情報学) / 甲第24938号 / 情博第849号 / 新制||情||142(附属図書館) / 京都大学大学院情報学研究科数理工学専攻 / (主査)准教授 原口 和也, 教授 山下 信雄, 教授 阿久津 達也 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

No description available.

No description available.

No description available.

No description available.

Quantitative Structure Activity Relationship (QSAR) is a very useful computa-tional method which has facilitated great progress in drug development [1]. Thismethod can be used to predict a molecule’s activity against a certain target justby comparing its structural characteristics (i.e., molecular descriptors) with thosebelonging to molecules of known activity. QSAR modeling is fueled by online freedatabases consisting of millions of active and inactive molecules and by MachineLearning (ML) Methods that enable data analysis. To ensure successful implemen-tation of ML models, there is a range of evaluation methods to estimate their perfor-mance and applicability domain. So far, a great deal of research has focused on theuse of Support Vector Machines (SVMs) to classify molecules with the use of theirMolecular Signature Fingerprints as descriptors [2]. However, another MachineLearning algorithm, Deep Neural Networks (DNNs), an improvement of single-layer Neural Networks, is rising in popularity in various fields including moleculeclassification. The two models were compared using CPSign software which intro-duces Conformal Prediction, to evaluate the reliability of model predictions basedon performance for individual compounds rather than mean performance on agiven test set. Three types of descriptors were used: Molecular Signature Finger-prints, Extended Connectivity Fingerprints and physicochemical descriptors. Thecomparison showed that Multilayer Perceptron (MLP) which was used as a DNNrepresentative in current context, had performance similar to the shallower SVMmodels but additionally demanded longer training times [3]. It can be concludedthat in the field of QSAR with the aforementioned descriptors, when the numberof examples used for training is not immense, Support Vector Machines might per-form equally well and demand less resources and time than the more sophisticated MLPs.