Background Pancreatic ductal adenocarcinoma (PDAC) is a fatal malignant neoplasm. It is necessary to improve the understanding of the underlying molecular mechanisms and identify the key genes and signaling pathways involved in PDAC. Methods The microarray datasets GSE28735, GSE62165, and GSE91035 were downloaded from the Gene Expression Omnibus. Differentially expressed genes (DEGs) were identified by integrated bioinformatics analysis, including protein–protein interaction (PPI) network, Gene Ontology (GO) enrichment, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses. The PPI network was established using the Search Tool for the Retrieval of Interacting Genes (STRING) and Cytoscape software. GO functional annotation and KEGG pathway analyses were performed using the Database for Annotation, Visualization, and Integrated Discovery. Hub genes were validated via the Gene Expression Profiling Interactive Analysis tool (GEPIA) and the Human Protein Atlas (HPA) website. Results A total of 263 DEGs (167 upregulated and 96 downregulated) were common to the three datasets. We used STRING and Cytoscape software to establish the PPI network and then identified key modules. From the PPI network, 225 nodes and 803 edges were selected. The most significant module, which comprised 11 DEGs, was identified using the Molecular Complex Detection plugin. The top 20 hub genes, which were filtered by the CytoHubba plugin, comprised FN1, COL1A1, COL3A1, BGN, POSTN, FBN1, COL5A2, COL12A1, THBS2, COL6A3, VCAN, CDH11, MMP14, LTBP1, IGFBP5, ALB, CXCL12, FAP, MATN3, and COL8A1. These genes were validated using The Cancer Genome Atlas (TCGA) and Genotype–Tissue Expression (GTEx) databases, and the encoded proteins were subsequently validated using the HPA website. The GO analysis results showed that the most significantly enriched biological process, cellular component, and molecular function terms among the 20 hub genes were cell adhesion, proteinaceous extracellular matrix, and calcium ion binding, respectively. The KEGG pathway analysis showed that the 20 hub genes were mainly enriched in ECM–receptor interaction, focal adhesion, PI3K-Akt signaling pathway, and protein digestion and absorption. These findings indicated that FBN1 and COL8A1 appear to be involved in the progression of PDAC. Moreover, patient survival analysis performed via the GEPIA using TCGA and GTEx databases demonstrated that the expression levels of COL12A1 and MMP14 were correlated with a poor prognosis in PDAC patients (p < 0.05). Conclusions The results demonstrated that upregulation of MMP14 and COL12A1 is associated with poor overall survival, and these might be a combination of prognostic biomarkers in PDAC.

中文翻译：

---

使用综合生物信息学分析确定 MMP14 和 COL12A1 作为胰腺导管腺癌预后生物标志物的潜在组合

背景胰腺导管腺癌（PDAC）是一种致命的恶性肿瘤。有必要提高对潜在分子机制的理解，确定参与 PDAC 的关键基因和信号通路。方法 微阵列数据集 GSE28735、GSE62165 和 GSE91035 从 Gene Expression Omnibus 下载。通过综合生物信息学分析鉴定差异表达基因 (DEG)，包括蛋白质-蛋白质相互作用 (PPI) 网络、基因本体论 (GO) 富集和京都基因和基因组百科全书 (KEGG) 通路富集分析。PPI 网络是使用相互作用基因检索搜索工具 (STRING) 和 Cytoscape 软件建立的。GO 功能注释和 KEGG 通路分析是使用数据库进行注释、可视化、和集成发现。Hub 基因通过基因表达谱交互分析工具 (GEPIA) 和人类蛋白质图谱 (HPA) 网站进行了验证。结果 三个数据集共有 263 个 DEG（167 个上调和 96 个下调）。我们使用 STRING 和 Cytoscape 软件建立 PPI 网络，然后识别关键模块。从 PPI 网络中，选择了 225 个节点和 803 条边。最重要的模块由 11 个 DEG 组成，是使用 Molecular Complex Detection 插件识别的。由 CytoHubba 插件过滤的前 20 个中枢基因包括 FN1、COL1A1、COL3A1、BGN、POSTN、FBN1、COL5A2、COL12A1、THBS2、COL6A3、VCAN、CDH11、MMP14、LTBP1、IGFBP5、ALB、CXCL12、FAP 、MATN3 和 COL8A1。这些基因使用癌症基因组图谱 (TCGA) 和基因型-组织表达 (GTEx) 数据库进行验证，随后使用 HPA 网站验证编码的蛋​​白质。GO分析结果表明，20个hub基因中最显着富集的生物学过程、细胞成分和分子功能术语分别是细胞粘附、蛋白质细胞外基质和钙离子结合。KEGG通路分析显示，20个hub基因主要富集于ECM-受体相互作用、粘着斑、PI3K-Akt信号通路和蛋白质消化吸收。这些发现表明 FBN1 和 COL8A1 似乎参与了 PDAC 的进展。而且，使用 TCGA 和 GTEx 数据库通过 GEPIA 进行的患者生存分析表明，COL12A1 和 MMP14 的表达水平与 PDAC 患者的不良预后相关（p < 0.05）。结论结果表明，MMP14 和 COL12A1 的上调与较差的总生存期相关，这可能是 PDAC 中预后生物标志物的组合。