MDR-ER: Balancing Functions for Adjusting the Ratio in Risk Classes and Classification Errors for Imbalanced Cases and Controls Using Multifactor-Dimensionality Reduction

Abstract

BackgroundDetermining the complex relationship between diseases, polymorphisms in human genes and environmental factors is challenging. Multifactor dimensionality reduction (MDR) has proven capable of effectively detecting statistical patterns of epistasis. However, MDR has its weakness in accurately assigning multi-locus genotypes to either high-risk and low-risk groups, and does generally not provide accurate error rates when the case and control data sets are imbalanced. Consequently, results for classification error rates and odds ratios (OR) may provide surprising values in that the true positive (TP) value is often small.Methodology/Principal FindingsTo address this problem, we introduce a classifier function based on the ratio between the percentage of cases in case data and the percentage of controls in control data to improve MDR (MDR-ER) for multi-locus genotypes to be classified correctly into high-risk and low-risk groups. In this study, a real data set with different ratios of cases to controls (1∶4) was obtained from the mitochondrial D-loop of chronic dialysis patients in order to test MDR-ER. The TP and TN values were collected from all tests to analyze to what degree MDR-ER performed better than MDR.Conclusions/SignificanceResults showed that MDR-ER can be successfully used to detect the complex associations in imbalanced data sets.

High Order Gene-Gene Interactions in Eight Single Nucleotide Polymorphisms of Renin-Angiotensin System Genes for Hypertension Association Study

“…In unbalanced function based MDR [ 25 ], the ratio between the percentages of cases in each genotype combination of case data (i.e., a cell in n × n grid for cases) to the percentages of controls in each genotype combination of control data (i.e., a cell in n × n grid for controls) is proposed to classify the data to the high- and low-risk groups. Thus, the highest ratio between case and control groups can be clearly detected.…”

“…Here, we describe a case-control study of hypertension susceptibility that specifically evaluates gene-gene interactions using unbalanced function based MDR [ 25 ] that combines traditional statistical methods with novel computational algorithms. The unbalanced function based MDR uses the ratio between the percentages of cases in each genotype combination of case data and the percentage of controls in each genotype combination of control data.…”

Breast cancer-associated high-order SNP-SNP interaction of CXCL12/CXCR4-related genes by an improved multifactor dimensionality reduction (MDR-ER)

“…MDR-ER was proposed previously (35). MDR-ER introduced the percentage concept to improve Equations 2 and 3 for imbalanced data.…”

“…However, MDR cannot quantitatively evaluate the disease susceptibility of genotype combinations (34). Recently, improved MDR methods such as MDR-ER (35) and weighted risk score-based MDR (34) were developed to solve this problem. Moreover, MDR-ER introduces two functions to improve the classification step and an evaluation of error rate in MDR, and it can be applied to the data set of imbalanced numbers of cases and controls.…”

An efficiency analysis of high-order combinations of gene–gene interactions using multifactor-dimensionality reduction

“…The first group contains modifications and combinations of biostatistics in MDR; this group includes entropy-based interpretation methods [ 11 ], the use of OR [ 12 ], generalized linear models [ 13 ], log-linear methods [ 14 ], Bayesian posterior probability [ 15 ], and model-based methods [ 16 ]. The second group focuses on particular data problems, such as imbalanced data [ 17 , 18 ], permutation testing [ 19 ], and missing data [ 20 ]. These extensions and modifications of MDR have been used to address different situations encountered in disease analysis.…”