Seed-weighted random walk ranking for cancer biomarker prioritisation: a case study in leukaemia.

A central focus of clinical proteomics for cancer is to identify protein biomarkers with diagnostic and therapeutic application potential. Network-based analyses have been used in computational disease-related gene prioritisation for several years. The Random Walk Ranking (RWR) algorithm has been successfully applied to prioritising disease-related gene candidates by exploiting global network topology in a Protein-Protein Interaction (PPI) network. Increasing the specificity and sensitivity ofbiomarkers may require consideration of similar or closely-related disease phenotypes and molecular pathological mechanisms shared across different disease phenotypes. In this paper, we propose a method called Seed-Weighted Random Walk Ranking (SW-RWR) for prioritizing cancer biomarker candidates. This method uses the information of cancer phenotype association to assign to each gene a disease-specific, weighted value to guide the RWR algorithm in a global human PPI network. In a case study of prioritizing leukaemia biomarkers, SW-RWR outperformed a typical local network-based analysis in coverage and also showed better accuracy and sensitivity than the original RWR method (global network-based analysis). Our results suggest that the tight correlation among different cancer phenotypes could play an important role in cancer biomarker discovery.