Computational phenotype prediction of ionizing-radiation-resistant bacteria with a multiple-instance learning model

Authors:
Sabeur Aridhi;Mondher Maddouri;Haitham Sghaier;Engelbert Mephu Nguifo
Affiliations:
Clermont University, Blaise and Pascal University, LIMOS, BP and University of Tunis El Manar, Tunis, Tunisia;University of Tunis El Manar, Tunis, Tunisia;Unit of Microbiology and Molecular Biology, National Center for Nuclear Sciences and Technologies (CNSTN), Sidi Thabet, Tunisia;Clermont University, Blaise and Pascal University, LIMOS, BP, Clermont-Ferrand, France
Venue:
Proceedings of the 12th International Workshop on Data Mining in Bioinformatics
Year:
2013

Citing 6
Cited 0

Solving the multiple instance problem with axis-parallel rectangles

Artificial Intelligence
A framework for multiple-instance learning

NIPS '97 Proceedings of the 1997 conference on Advances in neural information processing systems 10
Solving the Multiple-Instance Problem: A Lazy Learning Approach

ICML '00 Proceedings of the Seventeenth International Conference on Machine Learning
Predicting types of protein-protein interactions using a multiple-instance learning model

JSAI'06 Proceedings of the 20th annual conference on New frontiers in artificial intelligence
Data Mining: Concepts and Techniques

Data Mining: Concepts and Techniques
Feature extraction in protein sequences classification: a new stability measure

Proceedings of the ACM Conference on Bioinformatics, Computational Biology and Biomedicine

Quantified Score

Hi-index	0.00

Visualization

Abstract

Ionizing-radiation-resistant bacteria (IRRB) are important in biotechnology. The use of these bacteria for the treatment of radioactive wastes is determined by their surprising capacity of adaptation to radionuclides and a variety of toxic molecules. In silico methods are unavailable for the purpose of phenotypic prediction and genotype-phenotype relationship discovery. We analyze basal DNA repair proteins of most known proteomes sequences of IRRB and ionizing-radiation-sensitive bacteria (IRSB) in order to learn a classifier that correctly predicts unseen bacteria. In this work, we formulate the problem of predicting IRRB as a multiple-instance learning (MIL) problem and we propose a novel approach for predicting IRRB. We use a local alignment technique to measure the similarity between protein sequences to predict ionizing-radiation-resistant bacteria. The first results are satisfactory and provide a MIL-based prediction system that predicts whether a bacterium belongs to IRRB or to IRSB. The proposed system is available online.