Topological comparison method of multi-level biomedical interaction data for undiscovered public knowledge inference

Abstract

Since the increase of the public biomedical data, Undiscovered Public Knowledge (UPK, proposed by Swanson) became an important research topic. Many researchers tried to discover UPK, but these previous works required manual modulations to be applied to desired tasks, and had several inference limitations. In this paper, we propose TCM, Topological Comparison Method, to discover novel hypotheses using topological patterns of data. Topological patterns are connected sub-graphs of data, which store types of data, instead of values of data. TCM is appropriate for multi-level biomedical interaction data, because topological patterns are depending on the difference of types of entities and relations. By applying TCM to a public database, BIND, we could validate our method.