Bulgaria: Gene-disease link predictions accelerate drug discovery

Bulgarian company, Ontotext, part of Graphwise, is a global leader in graph technologies. It markets a semantic graph platform that integrates knowledge and disparate data into a holistic interface, supporting advanced retrieval-augmented generation and text and predictive analysis. These data-driven insights are advantageous for training artificial intelligence models to discover effective disease treatments, including the development of new drugs.

Ontotext appropriated its cutting-edge knowledge graph platform in the Eureka project FROCKG, collaborating with German partners Metaphacts and Paderborn University, and Swiss partner Zazuko. The consortium consolidated expertise in knowledge graphs, algorithms for fact veracity checking and graph path search to develop new knowledge graph embedding methods that help artificial intelligence predict gene-disease links. With this information, researchers and medical practitioners can explore disease markers in more depth and substantiate evidence about the effectiveness of new therapies.

The company’s commercial success has its roots in our dedication to championing research and innovation. Recent proof is the use of target discovery and its gene-disease link prediction method developed within a Eureka project by a top global early drug research and development centre to speed up their research on cancer therapies.

Zlatina Marinova – Vice President of Innovation at Graphwise

As a result of the FROCKG project, Graphwise expanded its portfolio of knowledge graph technologies, launching two new tools: Target Discovery and LinkedLifeData Inventory. These help pharmaceutical and biotechnology companies find new therapeutic targets and give integrated access to over 200 public datasets, respectively.

The research carried out in FROCKG was computationally intensive and time-consuming, typically posing a risk for companies investing resources without clear commercial benefits. However, support from Eureka and the Bulgarian National Innovation Fund gave Ontotext freedom to experiment with state-of-the-art methods for predicting gene-disease relationships. Ontotext also deepened its partnership with Metaphacts through collaboration on a demonstrator for knowledge graph management and fact curation, whilst working with Zazuko boosted its expertise in shapes constraint language rule validation.

With more than 50 completed European Union projects, Ontotext is the most prolific Bulgarian company in the international research, development and innovation landscape. In 2010, the Bulgarian Ministry of Education, Youth and Science awarded the company the Pythagoras Award for scientific developments benefitting society. FROCKG has reinforced Ontotext’s position, adding a competitive advantage to its products. Consequently, a top global early drug research and development centre is now using these solutions to focus its research and significantly speed up the development of novel cancer therapies. By leveraging artificial intelligence and expert-curated insights, the centre has prioritised the most promising cancer treatments and saved months of research time.

Today, Ontotext offers the most comprehensive and trusted industry solutions in the field of knowledge graphs and semantic technologies to clients worldwide. The development of a method for gene-disease link prediction delivers even more societal value: pre-clinical assessments for new treatments are accelerated tenfold and drug development risks are reduced.