Department of Bioinformatics, School of Biology and Basic Medical Sciences, Suzhou Medical College of Soochow University, 199 Renai Road, Suzhou, 215123, China.
Center for Systems Biology, Soochow University, 199 Renai Road, Suzhou, 215123, China.
J Transl Med. 2023 Mar 2;21(1):163. doi: 10.1186/s12967-023-04010-z.
Gastric cancer (GC) is a major cancer burden throughout the world with a high mortality rate. The performance of current predictive and prognostic factors is still limited. Integrated analysis is required for accurate cancer progression predictive biomarker and prognostic biomarkers that help to guide therapy.
An AI-assisted bioinformatics method that combines transcriptomic data and microRNA regulations were used to identify a key miRNA-mediated network module in GC progression. To reveal the module's function, we performed the gene expression analysis in 20 clinical samples by qRT-PCR, prognosis analysis by multi-variable Cox regression model, progression prediction by support vector machine, and in vitro studies to elaborate the roles in GC cells migration and invasion.
A robust microRNA regulated network module was identified to characterize GC progression, which consisted of seven miR-200/183 family members, five mRNAs and two long non-coding RNAs H19 and CLLU1. Their expression patterns and expression correlation patterns were consistent in public dataset and our cohort. Our findings suggest a two-fold biological potential of the module: GC patients with high-risk score exhibited a poor prognosis (p-value < 0.05) and the model achieved AUCs of 0.90 to predict GC progression in our cohort. In vitro cellular analyses shown that the module could influence the invasion and migration of GC cells.
Our strategy which combines AI-assisted bioinformatics method with experimental and clinical validation suggested that the miR-200/183 family-mediated network module as a "pluripotent module", which could be potential marker for GC progression.
胃癌（GC）是全球范围内主要的癌症负担，死亡率很高。目前预测和预后因素的表现仍然有限。需要进行综合分析，以确定准确的癌症进展预测生物标志物和预后生物标志物，以帮助指导治疗。
我们使用一种人工智能辅助的生物信息学方法，结合转录组数据和 microRNA 调控，来识别 GC 进展中关键的 microRNA 介导的网络模块。为了揭示该模块的功能，我们通过 qRT-PCR 对 20 个临床样本进行了基因表达分析，通过多变量 Cox 回归模型进行了预后分析，通过支持向量机进行了进展预测，并进行了体外研究以阐述其在 GC 细胞迁移和侵袭中的作用。
我们确定了一个稳健的 microRNA 调控网络模块，用于表征 GC 进展，该模块由七个 miR-200/183 家族成员、五个 mRNAs 和两个长非编码 RNA H19 和 CLLU1 组成。它们的表达模式和表达相关性模式在公共数据集和我们的队列中是一致的。我们的研究结果表明该模块具有双重生物学潜力：高风险评分的 GC 患者预后较差（p 值<0.05），并且该模型在我们的队列中预测 GC 进展的 AUC 值为 0.90。体外细胞分析表明，该模块可以影响 GC 细胞的侵袭和迁移。
我们的策略结合了人工智能辅助的生物信息学方法与实验和临床验证，表明 miR-200/183 家族介导的网络模块作为一种“多能模块”，可能是 GC 进展的潜在标志物。
