Application of personalized differential expression analysis in human cancer proteome

Owing to the recent technological advances, liquid chromatography-mass spectrometry (LC-MS)-based quantitative proteomics can measure expression of thousands of proteins from biological specimens. Currently, several studies have used the LC-MS-based proteomics to measure protein expression levels in human cancer. Identifying differentially expressed proteins (DEPs) between tumors and normal controls is a common way to investigate carcinogenesis mechanisms. However, most statistical methods used for DEPs analysis can only identify deregulated proteins at the population-level and ignore the heterogeneous differential expression of proteins in individual patients. Thus, to identify patient-specific molecular defects for personalized medicine, it is necessary to perform personalized differential analysis at the scale of a single sample. To date, there is a scarcity of systematic and easy-to-handle tool that could be used to evaluate the performance of individualized difference expression analysis algorithms in human cancer proteome. Herein, we developed a user-friendly tool kit, IDEP, to enable implementation and evaluation of personalized difference expression analysis algorithms. IDEP evaluates five rank-based tools (RankComp v1/v2, PENDA, Peng and Quantile) through classic computational and functional criteria in lung, gastric and liver cancer proteome. The results show that the within-sample relative expression orderings (REOs) of protein pairs in normal tissues were highly stable, which provided the basis for individual level DEPs analysis. Moreover, these individualized difference analysis tools could reach much higher efficiency in detecting sample-specific deregulated proteins than the group-based methods. Pathway enrichment and survival analysis results were dataset and analysis method dependent. In summary, IDEP has integrated necessary toolkits for individualized identification of DEPs and supported flexible methods evaluation analysis and visualization modules. It could provide a robust and scalable framework to extract personalized deregulation patterns and could also be used for the discovery of prognostic biomarkers for personalized medicine. ### Competing Interest Statement The authors have declared no competing interest.