One of the most important tasks of modern bioinformatics is the development of computational tools that can be used to understand and treat human disease. To date, a variety of methods have been explored and algorithms for predicting whether a protein is involved in disease are gaining in their utility. Here, we describe an algorithm for detecting protein-disease associations based on the human protein-protein interaction network, known gene-disease associations, protein sequence, and protein functional information at the molecular level. Our method, PhenoPred ("www.phenopred.org":www.phenopred.org), is supervised: first, we map each protein onto the spaces of disease and functional terms based on distance to all annotated proteins in the protein interaction network. We also encode sequence, function, physicochemical, and predicted structural properties, such as secondary structure and flexibility. We then train support vector machines to detect a protein’s disease function for a number of terms in Disease Ontology (DO). We provided evidence that, despite the noise/incompleteness of experimental data and unfinished ontology of diseases, identification of candidate genes and proteins can be successful even when a large number of candidate disease terms are predicted on simultaneously.