An Improved Information-Gain Approach for Epistasis Detection in Case-Control Study

作者：Liao, Zhongli; Zhu, Wen; Cai, Lijun^*

来源：Journal of Computational and Theoretical Nanoscience, 2015, 12(8): 1858-1863.

DOI：10.1166/jctn.2015.3970

• Epistasis Detection
• Information-Gain
• Case-Control Study

Epistasis is now believed to play important roles in complex human diseases. Many statistical methods have been proposed to detect and characterize epistatc interactions. In this paper, we developed a novel information-gain approach to measure the association between SNPs and disease status. We define a new relative information gain (RIG) and apply it to contingency tables of genotype combinations and disease status. The method betters take into account of the application in imbalanced datasets and add a penalty factor for the conditional entropy, which can better amplify the allele difference between case and control. The proposed method is verified in both simulated datasets and a real disease dataset.

• 出版日期2015-8
• 单位湖南大学