Pancreatic ductal adenocarcinoma (PDAC) is a malignant tumor with poor prognosis and high mortality. Although a large number of studies have explored its potential prognostic markers using traditional RNA sequencing (RNA-Seq) data, they have not achieved good prediction effect. In order to explore the possible prognostic signaling pathways leading to the difference in prognosis, we identified differentially expressed genes from one scRNA-seq cohort and four GEO cohorts, respectively. Then Cox and Lasso regression analysis showed that 12 genes were independent prognostic factors for PDAC. AUC and calibration curve analysis showed that the prognostic model had good discrimination and calibration. Compared with the low-risk group, the high-risk group had a higher proportion of gene mutations than the low-risk group. Immune infiltration analysis revealed differences in macrophages and monocytes between the two groups. Prognosis related genes were mainly distributed in fibroblasts, macrophages and type 2 ducts. The results of cell communication analysis showed that there was a strong communication between cancer-associated fibroblasts (CAF) and type 2 ductal cells, and collagen formation was the main interaction pathway.

胰腺导管腺癌（PDAC）是一种预后差、死亡率高的恶性肿瘤。尽管大量研究利用传统RNA测序（RNA-Seq）数据探索其潜在的预后标志物，但并未取得良好的预测效果。为了探索导致预后差异的可能的预后信号通路，我们分别从一个 scRNA-seq 队列和 4 个 GEO 队列中鉴定出差异表达的基因。 Cox和Lasso回归分析显示12个基因是PDAC的独立预后因素。 AUC和校准曲线分析表明该预后模型具有良好的判别性和校准性。与低风险组相比，高风险组的基因突变比例高于低风险组。免疫浸润分析揭示了两组之间巨噬细胞和单核细胞的差异。预后相关基因主要分布于成纤维细胞、巨噬细胞和2型导管中。细胞通讯分析结果表明，癌症相关成纤维细胞（CAF）和2型导管细胞之间存在很强的通讯，其中胶原蛋白形成是主要的相互作用途径。