Big Data Framework for Scalable and Efficient Biomedical Literature Mining in the Cloud

The massive size of available biomedical literature requires researchers to utilize novel big data technologies in data storage and analysis. Among them is cloud computing which has become the most popular solution for big data applications in industry. However, many bioinformaticians still rely on expensive and inefficient in-house infrastructure to discover knowledge from biomedical literature. Although some cloud-based solutions were constructed recently, they failed to sufficiently address a few key issues including scalability, flexibility, and reusability. Moreover, no study has taken computational cost into consideration. To fill the gap, we proposed a cloud-based big data framework that enables researchers to perform reproducible and scalable large-scale biomedical literature mining in an efficient and cost-effective way. Additionally, a cloud agnostic platform was constructed and then evaluated on two open access corpora with millions of full-text biomedical articles. The results indicate that our framework supports scalable and efficient large-scale biomedical literature mining.