Figure 4
Integrated interrogation of large-scale transcriptome and proteome data using SP3-CTP derived candidates highlights the differential biology of ovarian carcinoma histotypes.
From the analysis of a set of 18-tumour samples from unique individuals covering HGSC, CCC, and ENOC, histotypes were cross-compared to generate sets of enriched markers for each. (a) Boxplots depicting the RNA expression in the TCGA panel of HGSC ovarian cancer tumours of 5 proteins (FOLR1, CRIP1, MSLN, SNCG, CRABP2) found to have high, and 5 (LEFTY1, GDF15, QPCT, GPC3, CTH) with low expression in HGSC determined in the SP3-CTP proteomics data. Overlaid points indicate the expression of these proteins in each individual tumour of that type. The total number of tumours in the TCGA dataset used was 182. (b) Expression of the 10 SP3-CTP-derived ovarian cancer proteins in the IHC data taken from the HPA34. The HPA data (n = 12 total individuals) contains values for multiple ovarian carcinoma histotypes that are aggregated for this analysis. Expression values were calculated based on assigning a numerical score of 9 for ‘High’, 6 for ‘Medium’, 3 for ‘Low’, and 0 for ‘Not detected’. Each expression value was multiplied by the number of tumours assigned with that class, and the total sum calculated to generate a per protein expression estimate. (c) Expression of the 5 candidate HGSC proteins in MS-based proteomics data taken from the CPTAC study of ovarian cancer. Values are spectral counts calculated across all pools of samples taken from studies completed at the two CPTAC involved institutes (Supplemental Table 5).
