• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.76.2024.tde-05062024-092355
Document
Author
Full name
João Paulo Cassucci dos Santos
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Carlos, 2024
Supervisor
Committee
Bruno, Odemir Martinez (President)
Koide, Tie
Penna, Thadeu Josino Pereira
Title in Portuguese
Biologia de sistemas e aprendizagem de máquina: novas aplicações de métodos computacionais
Keywords in Portuguese
Aprendizagem de máquina
Bioinformática
Biologia de sistemas
Biologia molecular
Ciência de redes
Abstract in Portuguese
A ciência de redes nos permite modelar problemas multivariados e complexos de uma forma relativamente simples. Esta vantagem tem se demonstrado bastante promissora dentro do contexto de pesquisas interdisciplinares, pois ela permite caracterizar quantitativamente problemas que antes podiam apenas ser estudados qualitativamente. Um área promissora para a aplicação da ciência de redes é a da biologia molecular, em específico, a biologia de sistemas, onde o contexto em que elementos discretos estão inseridos importa mais do que suas propriedades isoladas. Nesta dissertação, buscamos explorar de duas maneiras distintas as propriedades de redes de modo a averiguar possíveis conclusões biológicas que podem ser extraídas a partir de diferentes experimentos biomoleculares. A primeira abordagem utiliza-se de um novo método de mensurar similaridade entre vetores conhecido como índice de coincidência, que demonstrou ser mais eficiente na extração de informação biológica em redes de interação enzima-enzima do que medidas de correlação tradicionalmente utilizadas para estas modelagens, como o r de Pearson e Spearman. A segunda abordagem aplica novos métodos de extração de características em redes complexas, como o Lifelike Network Automata e o Deterministic Tourist Walk, em conjunto com aplicações de algoritmos de aprendizagem de máquina para classificar bancos de dados de redes biológicas que poderão auxiliar na classificação de organismos e na predição de novas vias metabólicas.
Title in English
Systems biology and machine learning: new applications of computational methods
Keywords in English
Bioinformatics
Machine learning
Molecular biology
Network science
Systems biology
Abstract in English
Network Science allows us to model multivariate and complex problems in a relatively simple way. This advantage has been shown to be very promising in the context of interdisciplinary researches because it allows us to characterize problems quantitatively that before could only be studied qualitatively. One promising research area for the application of network science is molecular biology, in specific, systems biology, where the context in which the discrete elements belong is more important than their isolated properties. In this dissertation, we intended to explore in two distinct ways the network properties in order to investigate possible biological conclusions that can be extracted from different biomolecular experiments. The first approach uses a new way to measure similarity between vectors known as coincidence index, which was shown to be more effective in the extraction of biological information from enzyme-enzyme interaction networks than the more common correlation measurements traditionally used in these types of modelings, like Pearsons and Spearmans r. The second approach applies new complex network feature extraction techniques, such as Life-Like Network Automata and the Deterministic Tourist Walk, together with machine learning algorithms to classify biological networks datasets that can help in the classification of species and in the prediction of new metabolic pathways.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2024-06-05
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.