Case study

Using Contextual Causal Data To Explore Lung Cancer - Part 1

Introduction / situation

Worldwide there is a vast amount of data and scientific research being generated, with an estimated 57,700 peer-reviewed articles published a week. Oncology in particular is a prime area for research and development, as the second leading cause of death worldwide, accounting for an estimated 9.6 million deaths in 2018. Amongst cancer, lung cancer is the most common cancer worldwide and has one of the lowest survival rates. Therefore, it is at the forefront of scientific research and drug development. To further the advancements in treatments in oncology and scientific research it is important to understand the sum knowledge of the existing scientific evidence. Having a comprehensive view and understanding of the current research ensures a clear plan and robust hypothesis can be developed to challenge and supplement existing theories and support the path to clinical efficacy. However, databases and techniques currently used to explore the existing research typically lack context, includes data usually 6-months out of date and can be composed from a limited and biased set of sources.


Galactic AI™ can generate knowledge graphs instantly across an unlimited set of sources, providing reliable, accurate and thorough knowledge that can be used to investigate and develop complex biomedical hypotheses. Galactic AI™ uses advanced technology to automatically curate unstructured data from over 35 million documents from a variety of sources, including journal articles, clinical trials and grant information. As new sources are published valuable data around targets, diseases, biomarkers and more is automatically curated in-context to ensure there is no delay in valuable data appearing in structured biomedical databases.

Galactic AI™ uses a combination of artificial intelligence, deep learning and natural language in a process that has been refined over a decade. The platform automatically annotates scientific text to flag key concepts. Annotated concepts are derived from ontologies for example genes, proteins, chemicals and phenotypes. Galactic AI™ ontologies include existing ontologies as well as those updated and created by Biorelate to improve results. New ontologies and vocabularies can be built and added to the search to ensure a particular area is captured.

To understand these annotated concepts, Galactic AI™ analyses the interactions between them and determines the type of relationships they have with one another. To ensure accuracy and precision Galactic AI™ allocates a confidence score to the relationship. Each causal interaction is assigned additional context through the related concepts mentioned (e.g. tissue, disease, species etc.), with over 1.4 billion in-context datapoints identified. Context is subsequently used to build flexible knowledge graphs instantly under any desired criteria.

To build a lung cancer knowledge graph, 54 different lung-cancer related diseases were selected. The lung cancer knowledge graph generated by Galactic AI™ reveals over 71K distinct directed molecular interactions across 7.8K distinct proteins. There are a total of 202K protein-protein interactions and many other combinations across cells, phenotypes and other diseases mentioned in context.

Impact & Benefit


Galactic AI™ generates detailed content to create knowledge graphs swiftly and automatically, a task that would not be possible to conduct manually.


The speed at which biomedical literature is curated ensures the knowledge graph produced has the latest information and includes the most recent publications.


Knowledge graphs can be updated guaranteeing the data included is always up to date with the fast-moving pace of research.

Robust foundation

The knowledge graph is a powerful tool that can be adapted and used as a reliable foundation for further research or a deep-dive analysis.

Thank you for reading

Download this case study as PDF

How Can We Help You?

Get in touch with us to find out how we can transform your R&D.
Contact us