|
Abstract:
|
In systems biology, it is of great interest to identify previously unreported association of genes with various functions and diseases and relationships among these genes. Recently, biomedical literature has been considered as a valuable resource for this purpose but most approaches still focus on deterministic identification of direct relationships among genes. To address this limitation, we propose a computationally efficient biclustering approach which allows to identify indirect relationships among genes and facilitate their biological understanding from a text mining of biomedical literature. This approach also allows to integrate external biological knowledge as a prior information to guide the gene clustering, by using this information as soft constraints in order to take into account its incompleteness. We evaluate the proposed method with simulation studies and an application to studies of pathway-modulating genes in human, where we utilize biological pathway information available in existing databases as a prior knowledge.
|
