Protein Variability Patterns in Ovarian Serous Carcinoma

Abstract

Protein variability patterns in cancer reflects both technical variation and biological heterogeneity and may provide information beyond mean abundance changes. We analyzed protein-level coefficients of variation across four ovarian cancer proteomic datasets, comparing controls and cancer samples. We analyzed protein-level coefficients of variation (CVs) across four ovarian cancer proteomic datasets, comparing control and tumor samples. Among 7476 proteins, the median CV increased from 93% in controls to 126% in cancer, and the distributions differed significantly between the two groups (Wilcoxon p < 2.2 × 10−16; Kolmogorov–Smirnov p = 4.2 × 10−242). The largest increases in variability were observed among proteins with low inter-individual variability in controls, whereas proteins that were already highly variable showed more heterogeneous behavior, including decreases in CV. Thus, cancer was associated not only with an overall increase in variability but also with a redistribution of proteins across variability states. Gene Ontology analysis revealed functional differences between stable and highly variable proteins. Stable proteins were predominantly associated with intracellular, organelle-related, biosynthetic, and metabolic processes, whereas highly variable proteins were more frequently linked to membrane, vesicle-related, and signaling functions. Proteins that remain stable from the control to cancer state, as well as those that lose this stability during tumor development, may therefore be of particular interest. These results support protein variability as an additional analytical dimension alongside fold-change analysis for describing proteome instability and tumor heterogeneity. Inter-individual variation in protein abundance may reflect biological heterogeneity that is not captured by conventional comparisons of mean expression levels. Variability analysis complements conventional abundance-based approaches and provides an additional framework for characterizing tumor proteome heterogeneity.