A Queen’s University research lab in Canada wanted to expedite their analysis of how receptor tyrosine kinase (RET) mediated different processes in cancer. The lab used genome-wide large screens to identify candidates for validation, but the process was strenuous, requiring a lot of time and effort. It also created a huge backlog of over 2000 targets to be screened and validated, making it even more difficult to get results quickly and efficiently. To achieve better insights, the researchers needed to streamline their candidate validation process.
The lab’s main challenge was building a strong rationale for shortlisting candidates among thousands of genes. The manual process of individually validating these was time-intensive and resource-heavy. The lab had to screen for gene properties across different genomic databases, pathways, interactions, and literature. It was important to be able to provide solid evidence for the shortlisted candidates in terms of each of the properties and the consequences on pathological processes. The team shared that this process would take several months to complete.
The lab took advantage of the opportunity to use Ontotext’s Target Discovery Platform to streamline candidate validation and prioritization.
Target Discovery consumes data from Ontotext’s LinkedLifeData Inventory, which offers over 200 integration and analytics-ready datasets. They are maintained in a knowledge graph and can be plugged in and or taken out depending on specific needs. The unified access across multiple datasets makes it easy to find information quickly and efficiently all while linking facts, thus creating a sophisticated and enriched network of knowledge.
The platform uses these datasets to normalize the data found in scientific articles. Ontotext’s collection includes over 80 million articles, including scientific publications, patents, and clinical trials. The relevant information is automatically extracted with quality-proven AI models and integrated into the knowledge graph. The derived facts are then put into context with structured data, which provides clear evidence and provenance for all candidates. As a result, researchers get profound insights into the data stored in the knowledge graph and can mine long sequences of relationships.
Integrating all these tools helped researchers not only shortlist the set of genes most likely to succeed but also come up with a strong rationale for pursuing them.
When I think of Target Discovery, I believe it would be an integral part of every workflow. Now I always keep an open tab while I’m doing my work and it’s a super useful tool. It’s something I’d always refer to, whether that’s a literature search, looking for patterns I can use, or building a rationale for something. I can get the information I need and, within minutes, I’m able to build enough rationale to guide my research – Montdher Hussain, Mulligan Lab, Queen’s University.
Ontotext’s Target Discovery made it easy for researchers to leverage all proprietary and external resources they needed to answer their specific questions with high confidence. Rather than making users choose a ‘one style fits all’, the platform is very customizable and can adapt to different user needs with little effort. The scientists don’t require any technical knowledge or programming skills to mine complex biomedical relationships and get meaningful insights from all the data that is important to their work.
Our experience has been phenomenal. We initially thought we’d just plug in a set of genes, then get some results, and go and validate them. But it’s reached far beyond that. The platform has all these little tools that we can use to better prioritize candidates or build a stronger rationale – Montdher Hussain, Mulligan Lab, Queen’s University.