Extracting knowledge from literature to inform Tuberculosis drug and vaccine development
Develop a TB Knowledgebase, that interactively aggregates, organizes, and analyzes publicly available literature and clinical trial documentation to support researchers’ efforts challenged by the ever-growing body of TB literature.
What we did
The TB literature was mined to derive associative and causal relationships among host-, bacteria-, and intervention-related terms to provide a more efficient and effective way to identify knowledge, data, and models that can inform and guide model-informed drug and vaccine development. NLP-derived co-mention and relational networks were integrated with transcriptional profiles of M. tuberculosis and systems biology networks. Pathway databases, public preclinical and clinical data, and mathematical models were linked in the tool through COSBI’s text-mining infrastructure.
The user-friendly integrative TB Knowledgebase offers a visual exploration on more than 37,000 publications and 360 clinical trials integrated with pathophysiology conditions, bacterial genes and proteins, host immune responses, antibiotics, and vaccines. Visual clues, such as colors, shapes and edges, derived from the supporting evidence sentences, help the systems thinking approach to the integrated sources and involved entities to inform many aspects of drug and vaccine development, including study designs.
Azer K, Michelini S, Giampiccolo S, Parolo S, Leonardelli L, Lombardo R, Kaddi C, (2019) TB Knowledgebase: Interactive Application for Extracting Knowledge from the TB Literature to Inform TB Drug and Vaccine Development. The International Journal of Tuberculosis and Lung Disease: Abstract Book of 50th World Conference on Lung Health of the International Union Against Tuberculosis and Lung Disease (The Union) 22(11):S592.