MDR-ER: balancing functions for adjusting the ratio in risk classes and classification errors for imbalanced cases...

Area of Science:

• Genetics
• Bioinformatics
• Statistical genomics

Background:

• Complex relationships between diseases, gene polymorphisms, and environmental factors are challenging to determine.
• Multifactor dimensionality reduction (MDR) effectively detects epistasis but struggles with imbalanced datasets and accurate genotype classification.
• Traditional MDR may yield unreliable error rates and odds ratios with imbalanced case-control data.

Purpose of the Study:

• To introduce an improved MDR method (MDR-ER) that enhances the classification of multi-locus genotypes into high-risk and low-risk groups.
• To address the limitations of MDR in handling imbalanced case-control datasets.
• To improve the accuracy of error rates and odds ratio calculations in genetic association studies.

Main Methods:

• Developed a novel classifier function based on the ratio of case to control percentages within datasets.
• Applied the improved MDR method (MDR-ER) to a real-world dataset from chronic dialysis patients.
• Collected true positive (TP) and true negative (TN) values to compare MDR-ER performance against standard MDR.

Main Results:

• MDR-ER demonstrated successful classification of multi-locus genotypes in imbalanced datasets.
• The new method improved the accuracy of assigning genotypes to high-risk and low-risk categories.
• TP and TN values indicated superior performance of MDR-ER compared to standard MDR.

Conclusions:

• MDR-ER is a valuable tool for detecting complex genetic associations, particularly in datasets with imbalanced case and control groups.
• The method offers improved accuracy for genotype classification and risk assessment.
• This advancement facilitates more reliable analysis of genetic and environmental factors in disease.