SNP sets selection under mutual information criterion, application to F7/FVII dataset

H Brunel; A Perera; A Buil; M Sabater-Lleal; J C Souto; J Fontcuberta; M Vallverdu; J M Soria; P Caminal

doi:10.1109/IEMBS.2008.4650032

SNP sets selection under mutual information criterion, application to F7/FVII dataset

Annu Int Conf IEEE Eng Med Biol Soc. 2008:2008:3783-6. doi: 10.1109/IEMBS.2008.4650032.

Authors

H Brunel¹, A Perera, A Buil, M Sabater-Lleal, J C Souto, J Fontcuberta, M Vallverdu, J M Soria, P Caminal

Affiliation

¹ Institut de Bioenginyeria de Catalunya, Centre de Recerca en Enginyeria Biomédica, Departament de Enginyeria, de Sistemes, Automàtica i Informàtica industrial, Universitat Politcénica de Catalunya, Pau Gargallo 5, Barcelona, Spain. helena.brunel@upc.edu

PMID: 19163535
DOI: 10.1109/IEMBS.2008.4650032

Abstract

One of the main goals of human genetics is to find genetic markers related to complex diseases. In blood coagulation process, it is known that genetic variability in F7 gene is the most responsible for observed variations in FVII levels in blood. In this work, we propose a method for selecting sets of Single Nucleotide Polymorphisms (SNPs) significantly correlated with a phenotype (FVII levels). This method employs a feature selection algorithm (variant of Sequential Forward Selection, SFS) based on a criterion of statistical significance of a mutual information functional. This algorithm is applied to a sample of independent individuals from the GAIT project. Main SNPs found by the algorithm are in correspondence with previous results published using family-based techniques.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Algorithms
Artificial Intelligence
Cluster Analysis
Databases, Genetic
Factor VII / genetics*
Genomics / methods*
Humans
Models, Genetic
Models, Statistical
Models, Theoretical
Phenotype
Polymorphism, Single Nucleotide / genetics*

Substances

Factor VII