The overwhelming amount of biological data makes the assignment of candidate genes to diseases and biological pathways a formidable challenge. We present ENDEAVOUR, a generally applicable computational methodology to prioritize candidate genes based on their similarity to case-specific reference gene sets. Unlike previous methods, ENDEAVOUR is capable of flexibly utilizing multiple data sets from diverse sources. It allows the modular incorporation of de novo generated data sets and integrates distinct prioritizations into a global ranking by applying order statistics. We first validate the overallperformance in a statistical cross validation of 29 diseases and 3 biological pathways. We validate a novel candidate for DiGeorge syndrome in a zebrafish model and present several new candidates for congenital heart disease. We extend the basic ENDEAVOUR methodology using data from multiple species (human, mouse, rat, drosophila and C. elegans). We also present an alternative machine learning methodology for gene prioritization using kernel methods for novelty detection that outperforms our previous results.