X-meeting

André Fujita - University of São Paulo

        André Fujita

Habilitation (Livre-docência), 2016 - University of São Paulo, Brazil.
Title: Medidas de conectividade e métodos estatístico-computacionais para grafos com aplicações em Biologia Molecular e Neurociência.
Ph.D. in Science (Bioinformatics), 2007 - University of São Paulo, Brazil.
Title: Análise de dados de expressão gênica: normalização de microarrays e modelagem de redes regulatórias.
Advisor: Professor Carlos Eduardo Ferreira (Institute of Mathematics and Statistics, University of São Paulo)
Co-advisor: Professor Mari Cleide Sogayar (Chemistry Institute, University of São Paulo)
Fellowships: CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) and JICA (Japan International Cooperation Agency)
Bachelor in Computer Science, 2003 - University of São Paulo, Brazil.
Title: Anotação de genes associados com o controle da proliferação celular e a origem de tumores.
Advisor: Professor Mari Cleide Sogayar (Chemistry Institute, University of São Paulo)
Fellowship: CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico)


Abstract

Statistics on biological networks

The importance of statistics in biological sciences is increasing over the years. Besides, approaches based on systems biology, i.e., the modeling of the interactions among biological components (networks) is gaining much attention. However, biological networks present intrinsic fluctuations (random noise) that are not taken into account by classical methods grounded on graph theory. To construct a bridge between statistics and graphs, I will define graph spectral entropy and, based on it, present statistical methods to answer fundamental questions such as: Are two or more (sets of) biological networks equal (from a statistical perspective)? Are two vectors of biological networks correlated? How to estimate the parameters of biological network models? Which mathematical model best fits the biological network? Finally I will present a freely available open source software containing all the presented methods.