(a) An overview of the workflow employed in this study. Three iTRAQ replicates were carried out to ensure the consistency and reliability in results. (b) The protein molecular function classification of the 31 common proteins identified with similar iTRAQ trend across 3 iTRAQ replicates. The classification was performed using Panther classification systems (www.pantherdb.org/). A total of 40 assignments were obtained and sorted into 13 molecular classifications, despite the fact that there were only 31 proteins, as some of these proteins were assigned with more than 1 classification. The number showed in each category represented the number of protein assignment in a particular classification. The details for each classification are the following: Defense/immunity protein (Mbl2, Saa4, Apcs, C9); Extracellular matrix (Fga, Lrg1, Lum, Fgb, Igfals, Fgg); Hydrolase (Gpld1); Kinase (Egfr); Lyase (Car2); Miscellaneous function (Orm2); Molecular function unclassified (Apoc4); Receptor (Lifr, Egfr, Lrg1, Lum, Igfals); Select regulatory molecule (Itih3, Itih2, Serpina3k, Ambp, Itih1, Serpina1d); Signaling molecule (Fga, Egfr, Fgb, Fgg); Transfer/carrier protein (Hbb-b2, Apoc1, Saa4, Apod, Hbb-b1, Rbp4, Hba-a2;Hba-a); Transferase (F13a1); Transporter (Apoc1, Saa4)

(a) Representative immunoblot of ITIH3 validation. (b) Bar chart shows the average densitometry ratio obtained from the triplicate immunoblots of ITIH3 protein. The trends observed in the immunoblotting were similar to the protein expression level obtained via iTRAQ approach in which ITIH3 was more highly expressed in high tumor load when compared against control. (c) The overall protein profile of the pooled plasma samples from control, low, mid, and high tumor load mice. The gel was stained with SYPRO Ruby to ensure equal loadings during immunoblotting analysis.

(A) Triplicate immunoblots were carried out for each individual and their average densitometry reading were plotted in the bar chart according to their classification, i.e., normal and gastric cancer patients. The average densitometry reading for both the overall normal subject and gastric cancer group was also calculated and shown in the chart. The average expression level of ITIH3 in gastric cancer patients was found to be 1.9-fold higher compared to normal. (B) 2-sample t test analysis carried out using the average densitometry for each individual samples for the two categories. Significant difference (p-value <0.001) in ITIH3 expression level was observed between normal versus cancer groups. (C) The receiver operating characteristics (ROC) curve using ITIH3 average densitometry reading for all individual plasma samples (167) screened in the study. ITIH3 sensitivity and specificity in detecting gastric cancer were estimated to be 96% and 66% at the cutoff point of the average densitometry reading of 111 305. (D) A box-plot showing the distribution of ITIH3 expression within normal, early and late stage gastric cancer. Analysis of variance (ANOVA) analysis performed showed that a significant difference (p-value <0.001) in ITIH3 expression level was observed between normal versus early stage and late stage gastric cancer groups.

(A) An illustration of ITIH3 expression level in normal tissue, in which ITIH3 expression was found to be randomly expressed in discrete individual cell across the tissue, as shown by the arrows in the figure. (B) An example of ITIH3 expression in tumor tissue, where ITIH3 expression was widespread in many cancer cells. (C) ITIH3 expression level in a panel of gastric cancer cell lines, compared to the normal gastric cell line, HFE145. The dotted line in the figure highlighted ITIH3 expression level in normal gastric cell line HFE145.