Machine Learning to Establish Three Sphingolipid Metabolism Genes Signature to Characterize the Immune Landscape and Prognosis of Patients with Gastric Cancer

Background: Gastric cancer (GC) is one of the most common malignant tumors worldwide. Nevertheless, GC still lacks effective diagnosed and monitoring method and treating targets. This study used multi omics data to explore novel biomarkers and immune therapy targets around sphingolipids metabolism genes (SMGs).

Method: LASSO regression analysis was performed to filter prognostic and differently expression SMGs among TCGA and GTEx data. Risk score model and Kaplan-Meier were built to validate the prognostic SMG signature and prognostic nomogram was further constructed. The biological functions of SMG signature were annotated via multi omics. The heterogeneity landscape of immune microenvironment in GC was explored. qRT-PCR was performed to validate the expression level of SMG signature. Competing endogenous RNA regulatory network was established to explore the molecular regulatory mechanisms.

Result: 3-SMGs prognostic signature (GLA, LAMC1, TRAF2) and related nomogram were constructed combing several clinical characterizes. The expression difference and diagnostic value were validated by PCR data. Multi omics data reveals 3-SMG signature affects cell cycle and death via several signaling pathways to regulate GC progression. Overexpression of 3-SMG signature influenced various immune cell infiltration in GC microenvironment. RBP-SMGs-miRNA-mRNAs/lncRNAs regulatory network was built to annotate regulatory system.

Conclusion: Upregulated 3-SMGs signature are excellent predictive diagnosed and prognostic biomarkers, providing a new perspective for future GC immunotherapy.