Research On Serum Proteomics Of Hepatocellular Carcinoma Based On Data Independent Acquisition

Objectives Used Data independent acquisition(DIA)mass spectrometry(MS)to screen early diagnosis markers of hepatocellular carcinoma(HCC).Gene ontology(GO)and Proteinprotein interaction network(PPI)analysis was used to explore the relationship between biomarkers and the mechanism of development of HCC.Finally,used Parallel reaction monitoring(PRM),a targeted quantitative technology,to verify and evaluate the efficacy of HCC serum early diagnosis markers.Methods1.High performance liquid chromatography(HPLC)combining MARS-Human14 antibody column was used to separate high and low abundance proteins in 8 pools from320 serum samples.Polyacrylamide gel electrophoresis(SDS-PAGE)and Label-free quantification proteomics method were used to evaluate the separation effect of high and low abundance proteins.Ultra performance liquid chromatography(UPLC)continued to refine high and low abundance protein fractions,and collected MS data in Data dependent acquisition(DDA)mode.320 serum samples were used to prepare peptide samples directly to collect MS data in DIA mode.Integrate DDA and DIA data to generate HCC serum proteome MS library.2.320 DIA-MS data of serum samples were matched with the generated MS library to obtain serum proteomics data.Used P-value <0.05,Fold change(FC)≥1.2 or ≤0.83 as the threshold to screen HCC serum differential proteins,and used the time series expression pattern cluster analysis to select candidate markers in line with the trend of early diagnosis of HCC from them.Through GO and PPI analysis,we explored the correlation between candidates and the mechanism of occurrence and development of HCC.The importance of Randomforest model and Receiver operating characteristic curve(ROC)were used to evaluate the efficacy of ADGRG6 and optimize biomarker panel.3.Used Parallel reaction monitoring(PRM)to quantify the serum concentration of ADGRG6 in 75 samples of a validation cohort.Use ROC to test the area under the curve(AUC),sensitivity and specificity of ADGRG6 in the diagnosis of HCC.The efficacy was also tested in AFP-negative HCC to evaluate ADGRG6 combined with AFP in the diagnosis of HCC.Results1.SDS-PAGE showed that bands of low abundance protein(LAP)separated by the MARS-Human14 antibody column had darkened color,and Label-free MS analysis showed that the abundance and the number of LAPs were increased.The integrated serum proteome MS library contained 875 proteins.The quality of library was good,and proteins in it was in line with the nature and distribution of serum.123 proteins in the library had not been reported in Plasma proteome database(PPD).2.A total of 450 proteins could be identified in 320 serum samples,and 34 differentially changed proteins in HCC were screened out by differential analysis of samples from different groups.22 proteins were with the trend of early diagnosis of HCC.GO and PPI analysis showed that early diagnosis candidates were mainly related to the functions of the immune and metabolic regulation of HCC.The AUC of ADGRG6 in the diagnosis of HCC was 0.8398,the sensitivity was 0.7975,and the specificity was 0.8344.The AUC for the diagnosis of HCC in AFP-negative samples was 0.8210,the sensitivity was 0.8221,and the specificity was 0.8854.The AUC for diagnosis of HCC in DCP-negative samples was 0.8056,the sensitivity was 0.8000,and the specificity was 0.8880.The AUC of ADGRG6 combined with clinical AFP and DCP reached 0.9594,the sensitivity was0.9877,and the specificity was 0.8153.It had a good diagnostic efficiency for HCC.Randomforest machine learning model could optimize the HCC early diagnosis marker panel to the combination of 6 proteins.The AUC of panel predicting HCC was 0.990,the sensitivity was 0.980,and the specificity was 0.957.Panel had good HCC prediction performance.It could also effectively predict AFP-negative and DCP-negative HCC.3.After the quantitative verification of PRM,the concentration of ADGRG6 in HCC serum was significantly increased.The AUC of ADGRG6 in the diagnosis of HCC was0.7606,the sensitivity was 0.8980,and the specificity was 0.5385.ADGRG6 has diagnostic efficiency.The AUC of ADGRG6 combined with clinical AFP for diagnosis is0.9030,the sensitivity is 0.7708,and the specificity is 0.9231.Combined diagnosis can improve the efficiency.Conclusions1.MARS-Human 14 antibody column combined with UPLC to separate serum protein fractions can construct a serum proteome MS library efficiently.2.A HCC serum proteome expression profile is constructed.ADGRG6 and the panel composed of 6 proteins screened from it have good diagnostic performance for HCC.3.The concentration of ADGRG6 in HCC serum is significantly increased,which is a potential early diagnosis biomarker in HCC serum.ADGRG6 combines AFP to diagnose HCC can effectively improve the diagnostic efficiency.