Using the method of Multifactor Dimensionality Reduction (MDR) and its modifications for analysis of gene-gene and gene-environment interactions in genetic-epidemiological studies (review)
Background: In the genetic and epidemiological study of multifactorial signs (diseases), an important task is to assess the genetic and genetic-environmental interactions associated with the studied phenotype. The aim of the study: To carry out a systematic analysis of the data available in the modern literature on the possibilities of the method of Multifactor Dimensionality Reduction (MDR) and its various modifications (GMDR, MB-MDR) in the study of gene-gene and gene-environment interactions. Materials and methods: The review includes modern data of foreign and domestic articles on this topic found in Pubmed. Results: The MDR method makes it possible to evaluate gene-gene and gene-environment interactions associated with qualitative phenotypes, taking into account the correction for qualitative covariates and to carry out their validation using a permutation test. It also allows for cross-validation of models, assessment of the nature (synergy, additive, redundancy) and strength (contribution to entropy) of these interactions and their graphical visualization. This method makes it impossible to study quantitative phenotypes and to take into account quantitative covariates. The MB-MDR method allows to analyze the intergenic and gene-environment interactions associated with qualitative and quantitative phenotypes, to take into account covariates in the analysis, to validate the obtained models using the permutation test, and to determine individual combinations of factors associated with the studied phenotypes, taking into account covariates and significance (risk or protective value). The GMDR method makes it possible to evaluate gene-gene and gene-environment interactions associated with qualitative phenotypes with regard to correction for qualitative and quantitative covariates, to carry out their validation using the permutation test and visualize graphically; it allows for cross-validation of the most significant models, taking into account correction for covariates and multiple comparisons (permutation test). Conclusion: In the genetic-epidemiological study, the most optimal method is to use the MB-MDR method to establish the most significant SNP×SNP and gene-environment interactions, their validation by means of the permutation test, as well as to determine the specific combinations associated with the phenotype under study. Next, using the GMDR method, cross-validation of the most significant models, taking into account the correction for covariates and multiple comparisons (permutation test), and, finally, the use of the MDR method to estimate the nature (synergy, additive, redundancy) and strength (contribution to entropy) of SNP×SNP and gene-environment interactions and their graphical visualization.
Ponomarenko IV. Using the method of Multifactor Dimensionality Reduction (MDR) and its modifications for analysis of gene-gene and gene-environment interactions in genetic-epidemiological studies (review). Research Results in Biomedicine. 2019;5(1):4-21 (In Russian).
DOI: 10.18413/2313-8955-2019-5-1-0-1
While nobody left any comments to this publication.
You can be first.