Epistasis detection is vital to determining disease susceptibility in the human genome. With rapid advances in technology, multifactor dimensionality reduction (MDR) has become an effective algorithm for epistasis detection. Classification of high-risk (H) and low-risk (L) groups in MDR operations is a key topic, but it has not been thoroughly investigated. In this paper, we propose an improved fuzzy c-means-based entropy (FCME) approach to address the limitations of binary classification. For this approach, the degree of membership in MDR, referred to as FCMEMDR, was used. The FCME approach and MDR measure were integrated to enable more precise differentiation between similar frequencies of multifactor genotypes in the cases of possible epistasis. We used the MDR measures of correct classification rate and likelihood ratio. Numerous simulated datasets were applied, and the experimental results revealed two measures of FCMEMDR with higher detection rates than those of other MDR-based algorithms. Our analysis of binary and fuzzy classifications in MDR operations may offer insights into the problem of uncertainty in H/L classification. Two measures of FCMEMDR detected significant instances of epistasis associated with coronary artery disease in the Wellcome Trust Case Control Consortium dataset. FCMEMDR is freely available at https://gitlab.com/yudalinemail/fcmemdr.

中文翻译：

---

使用改进的基于模糊 C 均值的熵方法进行上位分析


上位性检测对于确定人类基因组中的疾病易感性至关重要。随着技术的快速进步，多因素降维（MDR）已成为上位检测的有效算法。 MDR操作中高风险（H）和低风险（L）群体的分类是一个关键话题，但尚未得到彻底研究。在本文中，我们提出了一种改进的基于模糊 c 均值的熵 (FCME) 方法来解决二元分类的局限性。对于这种方法，使用了 MDR 的隶属度，称为 FCMEMDR。 FCME 方法和 MDR 测量相结合，以便在可能上位的情况下，能够更精确地区分多因素基因型的相似频率。我们使用正确分类率和似然比的 MDR 度量。应用了大量的模拟数据集，实验结果表明 FCMEMDR 的两种测量方法比其他基于 MDR 的算法具有更高的检测率。我们对 MDR 操作中的二元和模糊分类的分析可以为 H/L 分类中的不确定性问题提供见解。 FCMEMDR 的两项测量在 Wellcome Trust 病例控制联盟数据集中检测到与冠状动脉疾病相关的显着上位性实例。 FCMEMDR 可在 https://gitlab.com/yudalinemail/fcmemdr 免费获取。