Department of Cardiology, Ningbo First Hospital, Ningbo 315000, Zhejiang Province, China.
Department of Cardiology, Ningbo Seventh Hospital, Ningbo 315000, Zhejiang Province, China.
Comput Biol Chem. 2017 Apr;67:194-199. doi: 10.1016/j.compbiolchem.2017.01.006. Epub 2017 Jan 19.
Our study was designed to identify the differential attractor modules related with hypertrophic cardiomyopathy (HCM) by integrating clustering-based on maximal cliques algorithm and Attract method.
We firstly recruited the HCM-related microarray data from ArrayExpress database. Next, protein-protein interaction (PPI) networks of normal and HCM were constructed and re-weighted using spearman correlation coefficient (SCC). Then, maximal cliques were found from the PPI networks through the clustering-based on maximal cliques approach. Afterwards, highly overlapped cliques were eliminated or merged according to the interconnectivity, and then modules were obtained. Subsequently, we used Attract method to identify differential attractor modules, following by the pathway enrichment analyses for genes in differential attractor modules.
After removing the cliques with nodes less than or equal to 4, 926 and 1118 maximal cliques in normal and HCM PPI networks were obtained for module analysis. Then, we obtained 32 and 55 modules from the PPI networks of normal and HCM, respectively. By comparing with normal condition, there were 5 module pairs with the same or similar gene composition. Significantly, based on attract method, we found that these 5 modules were differential attractors. Pathway enrichment analyses indicated that proteasome, ribosome and oxidative phosphorylation were the significant pathways.
Proteasome, ribosome and oxidative phosphorylation might play pathophysiological roles in HCM.
本研究旨在通过整合基于最大团的聚类算法和吸引方法，鉴定与肥厚型心肌病（HCM）相关的差异吸引子模块。
我们首先从 ArrayExpress 数据库中招募与 HCM 相关的微阵列数据。接下来，构建正常和 HCM 的蛋白质-蛋白质相互作用（PPI）网络，并使用 spearman 相关系数（SCC）对其进行重新加权。然后，通过基于最大团的聚类方法从 PPI 网络中找到最大团。之后，根据互连性消除或合并高度重叠的团，然后获得模块。随后，我们使用 Attract 方法识别差异吸引子模块，并对差异吸引子模块中的基因进行通路富集分析。
去除节点数小于或等于 4 的团后，正常和 HCM PPI 网络中分别获得了 926 个和 1118 个最大团进行模块分析。然后，我们分别从正常和 HCM 的 PPI 网络中获得了 32 个和 55 个模块。与正常情况相比，有 5 个模块对具有相同或相似的基因组成。值得注意的是，基于 attract 方法，我们发现这 5 个模块是差异吸引子。通路富集分析表明，蛋白酶体、核糖体和氧化磷酸化是显著的通路。
蛋白酶体、核糖体和氧化磷酸化可能在 HCM 中发挥病理生理作用。
