Loss-of-function screening using RNA interference or CRISPR approaches can be used to identify genes that specific tumor cell lines depend upon for survival. By integrating the results from screens in multiple cell lines with molecular profiling data, it is possible to associate the dependence upon specific genes with particular molecular features (e.g., the mutation of a cancer driver gene, or transcriptional or proteomic signature). Here, using a panel of kinome-wide siRNA screens in osteosarcoma cell lines as an example, we describe a computational protocol for analyzing loss-of-function screens to identify genetic dependencies associated with particular molecular features. We describe the steps required to process the siRNA screen data, integrate the results with genotypic information to identify genetic dependencies, and finally the integration of protein-protein interaction data to interpret these dependencies.