COVID-19 Knowledge Graphs

SARS-CoV-2 Molecular Relationship-Based Knowledge Graphs

Systemic analysis of existing large-scale biological and biomedical data is critical to develop new and effective treatment approaches against both complex and infectious diseases. In this context, “CROssBAR” has been developed, which is a comprehensive system that integrates large-scale biological and biomedical data from various resources and store them in a NoSQL database. In CROssBAR, integrated data is enriched with the deep-learning-based prediction of unknown relations, and the output data is represented as easy-to-interpret, interactive and heterogenous knowledge graphs that are displayed to users via an open-access web-service (

Using the CROssBAR system, COVID-19 knowledge graphs are constructed within the framework of host-pathogen molecular interactions and ontological relationships. COVID-19 knowledge graphs are expected to assist system-level research on relevant genes/proteins, drugs and drug candidate compounds that target these biomolecules, biological mechanisms, and other diseases involving these interactions (

Two different versions of COVID-19 knowledge graphs have been constructed, (i) a large-scale version that includes nearly the entirety of the related information on different integrated data sources, which is ideal for further network and/or machine learning based analysis ( (Figure 1), and (ii) a simplified version distilled to include only the curated molecular interactions (from UniProt-COVID-19 portal), which is ideal for a quick review and fast interpretation ( (Figure 2).

Şekil 1

Figure 1. COVID-19 large-scale knowledge graph.


Şekil 2

Figure 2. COVID-19 simplified knowledge graph.