The RIght TARGET
Human biology is one of the most complex information systems, and any variations of the underlying biological processes can cause symptoms and diseases to occur. Yet, the origin of a disease and the exact biological processes or pathways involved are often not clear. Determining which genes and proteins are associated with a disease and can be altered is the first step in being able to develop effective drug treatments. To do this, we bring together experts across ML, engineering, biology, bioinformatics, and chemistry together for each of our research programmes.
Technology enabled target identification.
We use machine learning and data science to guide the entire process of target identification. Relation inference AI models help us predict potential non-obvious disease targets that may be overlooked by scientists. Our differential expression based models help us identify proteins or genes that express differently in a disease and healthy cell.
The workflows we have developed look at the data in many different ways at the same time, allowing our scientists to establish solid hypotheses for possible drug targets.
Data-driven triage process.
Our AI triage process automatically orders targets based on criteria such as chemical opportunity, safety and druggability and then presents the ranked targets, together with rich metadata, to our scientists for an informed and nuanced evaluation. This allows them to focus on the most promising targets to progress into molecular design. These targets are tested in the lab and the results are fed back to our inference models.
Disease sprint for glioblastoma
The Multi-disciplinary approach to target identification.
Our team demonstrates the target identification process for Glioma Multiforme - a severe form of brain cancer - using machine learning methods. Their aim was to identify genes or proteins involved in the development of glioblastoma stem cells - a type of cell at the origin of the disease which is resistant to current treatments.