Covid19 Literature Browser For Scientific Investigations

Researchers around the world are in a race against the clock to find treatments, vaccines and diagnosis methods for COVID19. They can succeed only if they share and remain abreast of the latest developments concerning COVID19. A major hurdle lies in the necessity to digest an abundant scientific literature which grows by the day. Exploring large corpora comprising tens of thousands of publications about COVID19 is practically infeasible without appropriate visual text analytics software, combining the power of text mining algorithms with the flexibility of interactive visualizations. The present proposal combines two existing software assets from LIST and LCSB to address this problem. Papyrus is a web-based fully operational corpus visualization software providing an interactive visual map of topics extracted automatically without prior knowledge of the corpus. Unlike search engines or faceted search, Papyrus provides a user-friendly overview of the corpus and full drill-down capabilities into topics of interest. The user is typically able to home in on a few articles, usually 2-10 articles, addressing a specific question, within a couple of clicks. Going beyond the simple use of the MESH ontology currently achieved in Papyrus, we will reuse the BioKB software of LCSB to annotate the textual content of the corpora at hand. Medical researchers using Papyrus will be able to better distinguish the names of genes, proteins, chemicals, living organisms etc. thanks to this new synergy. Another major benefit of BioKB consists in the extraction of biologically meaningful relationships between bioentities occurring in the same sentence. We will enrich these relationships with non-biological entities encountered in the text, like geographic locations, author names, research institutions, the papers in which they are mentioned etc. We will support the exploratory analysis of such multiway relationships by combining tensor coclustering methods with interactive multilayer network visualizations. We would like to evaluate the usefulness of this approach for the discovery of subtle information by COVID19 researchers. The result of this integration will be accessible freely online at https://colibri.list.lu/ and will be advertised on COVID19 resource repositories to support researchers around the world and collect their feedback.