HitWalker: variant prioritization for personalized functional cancer genomics

Abstract

Summary: Determining the functional relevance of identified sequence variants in cancer is a prerequisite to ultimately matching specific therapies with individual patients. This level of mechanistic understanding requires integration of genomic information with complementary functional analyses to identify oncogenic targets and relies on the development of computational frameworks to aid in the prioritization and visualization of these diverse data types. In response to this, we have developed HitWalker, which prioritizes patient variants relative to their weighted proximity to functional assay results in a protein-protein interaction network. It is highly extensible, allowing incorporation of diverse data types to refine prioritization. In addition to a ranked list of variants, we have also devised a simple shortest path-based approach of visualizing the results in an intuitive manner to provide biological interpretation.

Availability and implementation: The program, documentation and example data are available as an R package from www.biodevlab.org/HitWalker.html.

Fig. 1.

Modified visualization output from HitWalker displaying the top three assay hits (EPHA4, JAK3 and FRK) and variants (FLT3, ZAK and PRKCE) for an acute myeloid leukaemia patient. Note that other hits are pulled out and annotated, as they are on the shortest path. Gene names are provided for each node. For nodes containing variants (blue), frequency information is reported in terms of the patient cohort counts (F), as well as the RWR rank (R). Red and green nodes indicate siRNA and gene target hits, respectively. Dotted borders indicate absence of capture probes for a given gene. Dashed borders indicate functional assay targets whose inhibition did not significantly alter cell viability. Confidence scores for the interactions between the two genes are reported near the lines connecting two given genes