We present a data integration and data mining platform, called BioGraph, for knowledge discovery in the biomedical domain. BioGraph allows for the automated formulation of comprehensible functional hypotheses relating a concepts to targets. A typical setting in which BioGraph can assist, is gene prioritization. That is, given the researcher's interest in a certain disease, predict those genes that are most likely of being involved in this disease. Our system is based on cutting-edge graph and network mining techniques, adapted to specific demands and properties of data and researchers in the Biomedical domain. The basis is constructed by the integration of heterogeneous biomedical knowledge bases. On this unified network BioGraph provides literature supported indirect functional relations. By assessing the plausibility and specificity of these hypothetical functional paths, the unsupervised methodology is capable of appraising and ranking of research targets, without requiring prior domain knowledge from the user. BioGraph is implemented as a web service and is available at: www.biograph.be.
展开▼