Motivation

Recently, multi-objective swarm intelligence optimization (SIO) algorithms have attracted considerable attention as disease model-free methods for detecting high-order single nucleotide polymorphism (SNP) interactions. However, a strict Pareto optimal set may filter out some of the SNP combinations associated with disease status. Furthermore, the lack of heuristic factors for finding SNP interactions and the preference for discrimination approaches to disease models are considerable challenges for SIO.

Method

In this study, we propose a multi-population harmony search (HS) algorithm dedicated to the detection of high-order SNP interactions (MP-HS-DHSI). This method consists of three stages. In the 1^st stage, HS with multi-population (multi-harmony memories) is used to discover a set of candidate high-order SNP combinations having an association with disease status. In HS, multiple criteria (Bayesian network-based K2-score, Jensen-Shannon (JS) divergence, likelihood ratio (LR) and normalized distance with joint entropy (ND-JE)) are adopted by four harmony memories to improve the ability to discriminate diverse disease models. A novel evaluation criterion named ND-JE is proposed to guide HS to explore clues for high-order SNP interactions. In the 2^nd and 3^rd stages, the G-test statistical method and multifactor dimensionality reduction (MDR) are employed to verify the authenticity of the candidate solutions, respectively.

Results

We compared MP-HS-DHSI with four state-of-the-art SIO algorithms for detecting high-order SNP interactions for 20 simulation disease models and a real dataset of age-related macular degeneration (AMD). The experimental results revealed that our proposed method can accelerate the search speed efficiently and enhance the discrimination ability of diverse epistasis models.

Availability and implementation

https://github.com/shouhengtuo/MP-HS-DHSI.

Supplementary information

Supplementary dataSupplementary data are available at Bioinformatics online.

中文翻译：

---

用于高阶SNP相互作用检测的多种群和谐搜索算法

动机

近年来，多目标群智能优化（SIO）算法作为检测高阶单核苷酸多态性（SNP）相互作用的无疾病模型方法而引起了广泛关注。但是，严格的帕累托最优集可能会过滤掉某些与疾病状态相关的SNP组合。此外，缺乏寻找SNP相互作用的启发性因素以及偏向于疾病模型的区分方法对于SIO来说是巨大的挑战。

方法

在这项研究中，我们提出了一种专用于检测高阶SNP相互作用（MP-HS-DHSI）的多种群和谐搜索（HS）算法。该方法包括三个阶段。在1^日阶段，HS具有多人口（多和谐存储器）用于发现一组具有与疾病状态关联的候选高阶SNP组合。在HS中，四个和声记忆采用了多个标准（基于贝叶斯网络的K2-得分，Jensen-Shannon（JS）发散，似然比（LR）和带联合熵的归一化距离（ND-JE））来提高区分各种疾病模型。提出了一种新的评价标准ND-JE，以指导HS探索高阶SNP相互作用的线索。在^第二和^第三 在这两个阶段中，分别采用G检验统计方法和多因素降维（MDR）来验证候选溶液的真实性。

结果

我们将MP-HS-DHSI与四种最新的SIO算法进行了比较，以检测20种模拟疾病模型和年龄相关性黄斑变性（AMD）的真实数据集的高阶SNP相互作用。实验结果表明，本文提出的方法可以有效地加快搜索速度，增强多种上皮模型的判别能力。

可用性和实施

https://github.com/shouhengtuo/MP-HS-DHSI。

补充资料

补充数据补充数据可从Bioinformatics在线获得。