Title
Mining nutrigenetics patterns related to obesity: use of parallel multifactor dimensionality reduction
Abstract
This paper aims to enlighten the complex etiology beneath obesity by analysing data from a large nutrigenetics study, in which nutritional and genetic factors associated with obesity were recorded for around two thousand individuals. In our previous work, these data have been analysed using artificial neural network methods, which identified optimised subsets of factors to predict one's obesity status. These methods did not reveal though how the selected factors interact with each other in the obtained predictive models. For that reason, parallel Multifactor Dimensionality Reduction pMDR was used here to further analyse the pre-selected subsets of nutrigenetic factors. Within pMDR, predictive models using up to eight factors were constructed, further reducing the input dimensionality, while rules describing the interactive effects of the selected factors were derived. In this way, it was possible to identify specific genetic variations and their interactive effects with particular nutritional factors, which are now under further study.
Year
DOI
Venue
2015
10.1504/IJBRA.2015.069194
International Journal of Bioinformatics Research and Applications
Field
DocType
Volume
Data mining,Nutrigenetics,Biology,Nutrigenomics,Multifactor dimensionality reduction,Curse of dimensionality,Obesity,Bioinformatics,Predictive modelling,Artificial neural network
Journal
11
Issue
Citations 
PageRank 
3
0
0.34
References 
Authors
5
4
Name
Order
Citations
PageRank
Katerina N. Karayianni100.34
Keith A. Grimaldi2131.67
Konstantina S. Nikita344866.23
Ioannis Valavanis49411.72