Reimagining the target selection process

Limited accessibility and usability of patient data impedes researchers’ ability to effectively utilize data in early drug discovery target selection. Better tools and collaboration throughout the R&D value chain can unlock new pathways to improving target prioritization, strengthening the art of target selection with more integration of real-world data. 

Selection of therapeutic targets is key to clinical success. By some measures, close to half of all clinical failures can be attributed to poor target selection—where either the targets were ineffective, or a link between the target and the disease had never actually been established. Improving target selection therefore holds promise for more successful clinical trial outcomes. But this requires tackling the challenge of drug development from the very start, by improving access to data, leveraging artificial intelligence, and enabling collaboration that combines technology with human ingenuity.

Lack of access to data can be a significant impediment during target selection. Historically, scientists rely on published literature and journal articles. However, inclusion of experimental data holds the promise to drive sound decision-making; but a lot of that data is locked away in siloes. Anything we can do to improve researchers’ access to clinical and real-world patient data will pave the way towards better and faster target selection and prioritization. 

But removing barriers to data access is just the first step. The amount and complexity of the available data—ranging from genomics and proteomics data to real-world patient data—is too much for any individual to grasp. Bringing sophisticated artificial intelligence tools and techniques to biologists and data scientists for collaboration in target selection and prioritization means that we can combine a deep understanding of biological processes with the ability to crunch the numbers. 

---

Digital transformation in the lab

LEARN MORE

---

Data, Collaboration, and Dashboards

Target ID combines data from a wide variety of sources—publicly available data, in-house data, proprietary data, and commercial data. The data is processed and then used to create and train AI models that help establish prioritization and rankings for different targets in the diseases of interest.

Research governance is also essential to enable collaboration and seamless workflows. Interpretation dashboards and graphical interfaces can help accelerate the work of the scientists by putting the right information into their hands at the right moment in time, in an intuitive and readily usable way. This enables visual assessments of whether additional data sources are needed, and other intuitive evaluations of data quality.

Designing based on Data

Every organization is different, with unique strengths and weaknesses and their own vision of what good target selection and target prioritization looks like. Prediction models and workflows need to be designed to cater to the diseases of interest and existing data sources for each company, which may include clinical trial and patient real-world data.

The future of target selection and prioritization includes much more data, of many different types and from more diverse sources. The challenge facing companies now is how to best prepare to manage and effectively utilize that data, with tools and processes that equip multi-disciplinary teams of scientists to harness and expand their best thinking. The last few years have seen a dramatic increase in data availability and the tools to understand it – a redesign of the target selection process to take advantage of these advances is possible and needed. 

<<< Start >>>

Jens Hoefkens 
PRINCIPAL DIRECTOR – Life Sciences 

I’d love to hear your thoughts on this topic.

<<< End >>>

Related content

<<< Start >>>

<<< End >>>

<<< Start >>>

<<< End >>>