Accelerating COVID-19 treatment innovation

Challenge: Unifying biomedical data to enable COVID-19 research
When COVID-19 research accelerated at an unprecedented pace, keeping up with the constant stream of new findings and integrating insights across different data sources, locations, and institutions became a daunting task for researchers. They needed a way to track the latest developments while also generating hypotheses faster and pushing treatment discovery forward.
Yet, patent data, genomics data, clinical trials, electronic health records, and research articles all came in different formats, scales, and standards. Integrating these diverse modalities into one coherent view was no small feat. On top of that, building a knowledge graph required careful calibration. If entities were modeled with too much detail, the system became inefficient; conversely, if too little detail was provided, critical insights risked being lost.
How we helped
We developed data integration and harmonization pipelines to bring different data sources into consistent formats and defined an ontology composed of entities of interest and their relationships. This data was ingested into a graph database (Neo4J) on a regular schedule, and we exposed a custom user interface to enable researchers to perform effective and intuitive queries on the resulting knowledge graph.
The impact: Transforming fragmented data into actionable biomedical insights
Our collaboration proved that the right data architecture and AI expertise can deliver measurable impact. Beyond immediate efficiency gains, the biomedical knowledge graph we built continues to empower research teams to innovate faster, make data-driven decisions, and reduce the time to scientific breakthroughs.
- Our solution required 10x fewer queries, which significantly reduced the effort needed to investigate a hypothesis and find reliable answers.
- We successfully unified diverse data sources, allowing for comprehensive analysis across gene data, clinical records, text, and images.
- We created a user-friendly tool to aid researchers in accelerating treatment innovation.

.png)
%20-%20main%20image%20(7)%20(1).avif)