Leveraging a Billion-Edge Knowledge Graph for Drug Re-purposing and Target Prioritization using Genomically-Informed Subgraphs

Abstract

Drug development is a resource and time-intensive process resulting in attrition rates of up to 90%. As a result, repurposing existing drugs with established safety and pharmacokinetic profiles is gaining traction as a way of accelerating therapeutics development. Here we have developed unique machine learning-driven Natural Language Processing and biomedical semantic technologies that mine over 53 million biomedical documents to automate the generation of a 911M edge knowledge graph. We then applied subgraph queries that relate drugs to diseases using genetic evidence to identify potential drug repurposing candidates for a broad range of diseases. We use Carney Complex, a disease with no known treatment, to illustrate our approach. This analysis revealed Ruxolitinib (Incyte, trade name Jakafi), a JAK1/2 inhibitor with an established safety and efficacy profile approved to treat myelofibrosis, as a potential candidate for the treatment of Carney Complex through off-target drug activity.

A large-scale evaluation of NLP-derived chemical-gene/protein relationships from the scientific literature: Implications for knowledge graph construction

“…GNBR's use of cooccurrence likely generates significant false positives as many sentences have cooccurring entities that have no semantic relationship. Modern NLP approaches will mitigate this challenge [11]. Row 4330 and 5594 exemplify how authors can express the consequence of 'inhibition', rather than stating it explicitly.…”

“…It is possible that a finer-grained relationship structure may be found, and coverage could be substantially increased due to fewer sentences being discarded. Other approaches such as Tellic's "Semantic Relationship Quantification" model-which scores the confidence in a meaningful sematic relationship between two entities in a sentence-could be used to check the accuracy of relationships in individual sentences [11].…”