Mining molecular interactions from scientific literature using cloud computing

Biomedical entities such as proteins and nucleic acids and its corresponding interactions is considered as one of the fundamental building block in understanding biological processes of organisms. Majority of this information is stored in an overwhelming amount of scientific literatures. Written in unstructured text format, this collection also employs natural language. Additionally, the demands for such extraction and analysis over a huge amount of biomedical abstracts and full texts requires effective means of data management and high performance computing systems. A key in efficiently acquiring this information involves an intelligent and effective system of information extraction using techniques in natural language processing, information retrieval and knowledge management. With this regard, we developed a straight-forward biomolecular mining system that recognizes the biological named entities, simple heuristic pattern rules in identifying molecular interactions and utilization of high performance cloud computing technology.