Combined metabolomics with transcriptomics reveals important serum biomarkers correlated with lung cancer proliferation through calcium signaling pathway

Lung cancer (LC) is still one of the most deadly cancers in the world, but there is a lack of non-invasive biomarkers to assist diagnosis and guide treatment. Herein, untargeted serum metabolomics of 57 LC patients and 59 HC people was performed based on gas chromatography-mass spectrography (GC-MS). Differential metabolites were identified by NIST14 database and human metabolome database (HMDB). The random forest (RF) classification algorithm was then carried out for selecting the optimal panel of biomarkers among differential metabolites. The biomarker panel was finally validated by support vector machine (SVM) and partial least squares-discriminant analysis (PLS-DA) classification algorithm. After that, each biomarker was targeted quantitated using GC-MS/SIM in a test set consisted of 50 LC patients and 41 HC people. Subsequently, the Network analysis of Ingenuity Pathway Analysis (IPA) was applied to browse the biomarker panel-related genes, and the transcriptome data from GEO database were used to verify the gene expression. In the results, five metabolites including cholesterol, oleic acid, myo-inositol, 2-hydroxybutyric acid and 4-hydroxybutyric acid were combined to be biomarker panel from totally 15 differential metabolites. The genes of INPP5A, MAP2K2, CAMK1D, GABARAPL1, NDUFS4/8 were involved in biomarker panel-related biological pathway, and calcium signaling pathways might mediate their role in the proliferation of LC. Our study provided effective candidate biomarkers for the diagnosis of LC, and elucidated part of the underlying pathological mechanism of LC from the perspective of metabolites.