Gastric cancer (GC) is a product of multiple genetic abnormalities, including genetic and epigenetic modifications. This study aimed to integrate various biomolecules, such as miRNAs, mRNA, and DNA methylation, into a genome-wide network and develop a nomogram for predicting the overall survival (OS) of GC.

A total of 329 GC cases, as a training cohort with a random of 150 examples included as a validation cohort, were screened from The Cancer Genome Atlas database. A genome-wide network was constructed based on a combination of univariate Cox regression and least absolute shrinkage and selection operator analyses, and a nomogram was established to predict 1-, 3-, and 5-year OS in the training cohort. The nomogram was then assessed in terms of calibration, discrimination, and clinical usefulness in the validation cohort. Afterward, in order to confirm the superiority of the whole gene network model and further reduce the biomarkers for the improvement of clinical usefulness, we also constructed eight other models according to the different combinations of miRNAs, mRNA, and DNA methylation sites and made corresponding comparisons. Finally, Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were also performed to describe the function of this genome-wide network.

A multivariate analysis revealed a novel prognostic factor, a genomics score (GS) comprising seven miRNAs, eight mRNA, and 19 DNA methylation sites. In the validation cohort, comparing to patients with low GS, high-GS patients (HR, 12.886; P < 0.001) were significantly associated with increased all-cause mortality. Furthermore, after stratification of the TNM stage (I, II, III, and IV), there were significant differences revealed in the survival rates between the high-GS and low-GS groups as well (P < 0.001). The 1-, 3-, and 5-year C-index of whole genomics-based nomogram were 0.868, 0.895, and 0.928, respectively. The other models have comparable or relatively poor comprehensive performance, while they had fewer biomarkers. Besides that, DAVID 6.8 further revealed multiple molecules and pathways related to the genome-wide network, such as cytomembranes, cell cycle, and adipocytokine signaling.

We successfully developed a GS based on genome-wide network, which may represent a novel prognostic factor for GC. A combination of GS and TNM staging provides additional precision in stratifying patients with different OS prognoses, constituting a more comprehensive sub-typing system. This could potentially play an important role in future clinical practice.

胃癌 (GC) 是多种遗传异常的产物，包括遗传和表观遗传修饰。本研究旨在将各种生物分子（如 miRNA、mRNA 和 DNA 甲基化）整合到全基因组网络中，并开发用于预测 GC 总生存期 (OS) 的列线图。

从癌症基因组图谱数据库中筛选出总共 329 例 GC 病例，作为训练队列，其中随机包含 150 个样本作为验证队列。基于单变量 Cox 回归和最小绝对收缩和选择算子分析的组合构建了全基因组网络，并建立了列线图来预测训练队列中的 1 年、3 年和 5 年 OS。然后在验证队列中根据校准、辨别和临床有用性评估列线图。之后，为了确认全基因网络模型的优越性，进一步减少生物标志物以提高临床实用性，我们还根据miRNA、mRNA、DNA甲基化位点的不同组合构建了其他8个模型，并进行了相应的比较。 . 最后，

多变量分析揭示了一种新的预后因素，即基因组学评分 (GS)，包括 7 个 miRNA、8 个 mRNA 和 19 个 DNA 甲基化位点。在验证队列中，与低 GS 患者相比，高 GS 患者（HR，12.886；磷< 0.001) 与全因死亡率增加显着相关。此外，在对 TNM 分期（I、II、III 和 IV）进行分层后，高 GS 组和低 GS 组之间的生存率也存在显着差异。磷< 0.001)。基于全基因组的列线图的 1 年、3 年和 5 年 C 指数分别为 0.868、0.895 和 0.928。其他模型具有可比或相对较差的综合性能，而它们的生物标志物较少。除此之外，DAVID 6.8 进一步揭示了与全基因组网络相关的多个分子和途径，例如细胞膜、细胞周期和脂肪细胞因子信号传导。

我们成功开发了一种基于全基因组网络的 GS，这可能代表了 GC 的一种新的预后因素。GS 和 TNM 分期的组合为具有不同 OS 预后的患者分层提供了额外的精确度，构成了一个更全面的子分型系统。这可能在未来的临床实践中发挥重要作用。