Detection of functional gene-gene interactions from observational gene expression data using classifier ensembles and Metalearning

• Grant number: 14/10852-7
• Support Opportunities: Scholarships abroad - Research Internship - Post-doctor
• Start date: September 01, 2014
• End date: March 31, 2015
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: André Carlos Ponce de Leon Ferreira de Carvalho
• Grantee: Edwin Rafael Villanueva Talavera
• Supervisor: Rainer Spang
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Institution abroad: University of Regensburg, Germany
• Associated to the scholarship: 12/22295-0 - Meta-learning and Bayesian Networks to Construct Classifier Ensembles for Gene-Expression Data, BP.PD

Abstract

Predict whether or not a gene will respond to the intervention of another gene only from their freely observed expression profiles is of great value in medical research. However, this is a challenging task because the randomness of the biological systems, the elevated noise levels of the data collection technologies and the high dimensionality involved. In this project we propose to apply machine learning (ML) approaches to learn the subtle differences in observational gene-expression patterns that could characterize gene-gene causal interactions. Specifically, we will study ways to take advantage of two promising approaches: Ensemble of Classifiers (EoC) and Metalearning. EoC have often shown their superiority in relation to single methods. Metalearning studies how learning algorithms could increase their efficiency through experience, having the potential to transform how ML methods are employed, although it has been little explored in bioinformatics applications. We believe that by properly joining the mentioned approaches we can obtain classification system that can accurately identify patterns of functional gene-gene interactions from observational gene expression data. (AU)