Towards automatic extraction of social networks of organizations in PubMed abstracts

Social network analysis (SNA) of organizations can attract great interest from government agencies and scientists for its ability to boost translational research and accelerate the process of converting research to care. For SNA of a particular disease area, we need to identify the key research groups in that area by mining the affiliation information from PubMed. This not only involves recognizing the organization names in the affiliation string, but also resolving ambiguities to identify the article with a unique organization. We present here a process of normalization that involves clustering based on local sequence alignment metrics and local learning based on finding connected components. We demonstrate the application of the method by analyzing organizations involved in angiogenensis treatment, and demonstrating the utility of the results for researchers in the pharmaceutical and biotechnology industries or national funding agencies.