Knowledge Discovery on Biomedical Literature:Validating and Quality Control on Cause and Effect Networks

Today the biomedical field mostly relies on systems biology
approaches such as integrative knowledge graphs to decipher mechanism of a disease, by considering system as a whole which is considered as a holistic approach. In that, disease modeling and pathway databases play
an important role. Knowledge Graphs built using Biological Expression Language (BEL, see is widely applied in biomedical domain to convert unstructured textual knowledge into a computable form. In addition, several information systems have been introduced to support curators generating these networks. We face several challenges while converting knowledge from literature into knowledge graphs. Here, we propose statistic measures based on document clustering to quantify completeness and coverage, to prove the quality of a knowledge graph
by identifying the scope, to distinguish and prioritize well-known, novel and missing knowledge based on literature.