PRER: A patient representation with pairwise relative expression of proteins on biological networks

Changes in protein and gene expression levels are often used as features in predictive modeling such as survival prediction. A common strategy to aggregate information contained in individual proteins is to integrate the expression levels with the biological networks. In this work, we propose a novel patient representation where we integrate proteins’ expression levels with the protein-protein interaction (PPI) networks: Patient representation with PRER (Pairwise Relative Expressions with Random walks) (PRER). PRER captures the dysregulation patterns of proteins based on the neighborhood of a protein in the PPI network. Specifically, PRER computes a feature vector for a patient by comparing the source protein’s expression level with other proteins’ levels that are within its neighborhood. The neighborhood of the source protein is derived by biased random-walk strategy on the network. We test PRER’s performance in survival prediction task in 10 different cancers using random forest survival models. PRER yields a statistically significant predictive performance in 9 out of 10 cancers when compared to the same model trained with features based on individual protein expressions. Furthermore, we identified the pairs of proteins that their interactions are predictive of patient survival but their individual expression levels are not. The set of identified relations provides a valuable collection of protein biomarkers with high prognostic value. PRER can be used for other complex diseases and prediction tasks that use molecular expression profiles as input. PRER is freely available at: https://github.com/hikuru/PRER.Author summary: Cancer remains to be one of the most prevalent and challenging diseases to treat. Cancer is a complex disease with several disrupted molecular mechanisms at play. The protein expression level is a fundamental indicator of how the molecular mechanisms are altered in each tumor. Predicting patient survival based on the changes is essential for understanding the cancer mechanisms and arriving at patient-specific treatment plans. For this task, existing machine learning models are used, such as random forest survival, which requires a feature-based representation of each patient based on her tumors. Most of these models use the individual molecular quantities of the tumors. However, cancer is a complex disease in which molecular mechanisms are dysregulated in various ways. In this work, we present a new patient representation scheme in which we integrate each tumor’s protein expression levels with their neighboring proteins’ expression levels in a protein-protein interaction network to capture patient-specific dysregulation patterns. Our results suggest that proteins’ relative expressions are more predictive than their individual expressions. We also analyze which of the protein interactions are more predictive of patient survival. The identified set of important protein interactions can be potentially used for cancer prognosis.