COSMONET: An R Package for Survival Analysis Using Screening-Network Methods

Identifying relevant genomic features that can act as prognostic markers for buildingpredictive survival models is one of the central themes in medical research, affecting the future ofpersonalized medicine and omics technologies. However, the high dimension of genome-wide omicdata, the strong correlation among the features, and the low sample size significantly increase thecomplexity of cancer survival analysis, demanding the development of specific statistical methodsand software. Here, we present a novel R package, COSMONET (COx Survival Methods based OnNETworks), that provides a complete workflow from the pre-processing of omics data to the selectionof gene signatures and prediction of survival outcomes. In particular, COSMONET implements (i) threedifferent screening approaches to reduce the initial dimension of the data from a high-dimensionalspace p to a moderate scale d, (ii) a network-penalized Cox regression algorithm to identify the genesignature, (iii) several approaches to determine an optimal cut-off on the prognostic index (PI) toseparate high- and low-risk patients, and (iv) a prediction step for patients' risk class based on theevaluation of PIs. Moreover, COSMONET provides functions for data pre-processing, visualization,survival prediction, and gene enrichment analysis. We illustrate COSMONET through a step-by-step Rvignette using two cancer datasets.