Abstract

The differential expression and direct targeting of mRNA by miRNA are two main logics of the traditional approach to constructing the miRNA-mRNA network. This approach, could be led to the loss of considerable information and some challenges of direct targeting. To avoid these problems, we analyzed the rewiring network and constructed two miRNA-mRNA expression bipartite networks for both normal and primary prostate cancer tissue obtained from PRAD-TCGA. We then calculated beta-coefficient of the regression-model when miR was dependent and mRNA independent for each miR and mRNA and separately in both networks. We defined the rewired edges as a significant change in the regression coefficient between normal and cancer states. The rewired nodes through multinomial distribution were defined and network from rewired edges and nodes was analyzed and enriched. Of the 306 rewired edges, 112(37%) were new, 123(40%) were lost, 44(14%) were strengthened, and 27(9%) weakened connections were discovered. The highest centrality of 106 rewired mRNAs belonged to PGM5, BOD1L1, C1S, SEPG, TMEFF2, and CSNK2A1. The highest centrality of 68 rewired miRs belonged to miR-181d, miR-4677, miR-4662a, miR-9.3, and miR-1301. SMAD and beta-catenin binding were enriched as molecular functions. The regulation was a frequently repeated concept in the biological process. Our rewiring analysis highlighted the impact of β-catenin and SMAD signaling as also some transcript factors like TGFB1I1 in prostate cancer progression. Altogether, we developed a miRNA-mRNA co-expression bipartite network to identify the hidden aspects of the prostate cancer mechanism, which traditional analysis -like differential expression- was not detect it.

摘要

miRNA的差异表达和直接靶向mRNA是构建miRNA-mRNA网络的传统方法的两个主要逻辑。这种方法可能会导致大量信息的丢失以及直接瞄准的一些挑战。为了避免这些问题，我们分析了重连网络，并为从 PRAD-TCGA 获得的正常和原发性前列腺癌组织构建了两个 miRNA-mRNA 表达二分网络。然后，我们计算了当 miR 依赖而 mRNA 独立时，每个 miR 和 mRNA 分别在两个网络中回归模型的 beta 系数。我们将重新连接的边缘定义为正常状态和癌症状态之间回归系数的显着变化。定义了通过多项分布重新连线的节点，并分析和丰富了重新连线边缘和节点的网络。在 306 个重新连线的边缘中，112 个（37%）是新的，123 个（40%）丢失，44 个（14%）被加强，27 个（9%）弱连接被发现。106 个重连 mRNA 的中心度最高的是 PGM5、BOD1L1、C1S、SEPG、TMEFF2 和 CSNK2A1。68 个重连 miR 中中心性最高的是 miR-181d、miR-4677、miR-4662a、miR-9.3 和 miR-1301。SMAD 和 β-连环蛋白结合作为分子功能得到丰富。调节是生物过程中经常重复的概念。我们的重连分析强调了 β-连环蛋白和 SMAD 信号传导以及一些转录因子（如 TGFB1I1）在前列腺癌进展中的影响。共，