Mining nutrigenetics patterns related to obesity: use of parallel multifactor dimensionality reduction

Area of Science:

• Nutrigenetics
• Obesity Etiology
• Computational Biology

Background:

• Obesity is a complex condition with multifactorial origins.
• Previous research utilized artificial neural networks to identify predictive factors for obesity status.
• Understanding the interactions between genetic and nutritional factors is crucial for elucidating obesity's etiology.

Purpose of the Study:

• To investigate the interactive effects of pre-selected nutrigenetic factors on obesity.
• To identify specific genetic variations and their interplay with nutritional factors.
• To derive rules describing these interactive effects using parallel Multifactor Dimensionality Reduction (pMDR).

Main Methods:

• Analysis of a large nutrigenetics study dataset comprising approximately two thousand individuals.
• Application of parallel Multifactor Dimensionality Reduction (pMDR) to analyze pre-selected factor subsets.
• Construction of predictive models using up to eight factors and derivation of interaction rules.

Main Results:

• pMDR successfully reduced input dimensionality while identifying significant interactions.
• Specific genetic variations were found to interact with particular nutritional factors.
• Rules describing these interactive effects were successfully derived.

Conclusions:

• The study successfully identified complex interactions between genetic and nutritional factors in obesity.
• These findings provide a deeper understanding of obesity's etiology.
• The identified interactions warrant further investigation for potential clinical applications.