Development of Prognostic Signature Based on Pan-cancer Proteomics

Abstract

Background: Utilizing genomic data to predict cancer prognosis was insufficient. Proteomics can improve our understanding of the etiology and progression of cancer and improve the assessment of cancer prognosis. Based on CPTAC (Clinical Proteomic Tumor Analysis Consortium) which has generated extensive proteomics data of the vast majority of tumors, we can perform a proteomic pan-carcinoma analysis.Methods: The proteomics data and clinical features of cancer patients were collected from CPTAC. We screened 69 differentially expressed proteins with R software. GO and KEGG analysis were performed to clarify the function of these proteins. The DEPs-based prognostic model was identified by least absolute shrinkage and selection operator (LASSO)-Cox regression model. The time-dependent receiver operating characteristics analysis was used to evaluate the ability of the prognostic model to predict overall survival.Results: A total of 69 differentially expressed proteins were screened in five different types of cancers: hepatocellular carcinoma (HCC), lung adenocarcinoma (LUAD), children's brain tumor tissue consortium (CBTTC), clear cell renal cell carcinoma (CCRC) and uterine corpus endometrial carcinoma (UCEC). Furthermore, the differentially expressed proteins were related to cell metabolism, cell proliferation and extracellular matrix. Then 24 DEPs-based classifiers for predicting OS was developed by LASSO-Cox regression model in training cohort, which was validated in validation cohort. Conclusions: In the present study, we identified DEPs-based survival-predictor model to predict most cancers. We are the first group to utilize proteomics to construct a pan-cancer prognosis model, which could accurately and effectively predict the survival rate of most cancers.