Evaluation of single-sample network inference methods for precision oncology
A major challenge in precision oncology is to identify targetable cancer vulnerabilities in individual patients. Modelling high-throughput omics data in biological networks allows identifying key molecules and processes of tumorigenesis. Traditionally, network inference methods rely on many samples to contain sufficient information for learning and predicting gene interactions for a group of patients. However, to implement patient-tailored approaches in precision oncology, we need to interpret omics data at the level of the individual patient. Several single-sample network inference methods have been developed that infer biological networks for an individual sample from bulk RNA-seq data. However, only a limited comparison of these methods has been made. Moreover, many methods rely on normal tissue samples as reference point for the tumor samples, which is not always available. Here, we conducted an evaluation of the single-sample network inference methods SSN, LIONESS, iENA, CSN and SSPGI using expression profiles of lung and brain cancer cell lines from the CCLE database. The methods constructed networks with distinct network topologies, as observed by edge weight distributions and other network characteristics. Further, hub gene analyses revealed different degrees of subtype-specificity across methods. Single-sample networks were able to distinguish between tumor subtypes, as exemplified by edge weight clustering, enrichment of known subtype-specific driver genes among hub gene sets, and differential node importance. Finally, we show that single-sample networks correlate better to other omics data from the same cell line as compared to aggregate networks. Our results point to the important role of single-sample network inference in precision medicine.