In early stage drug discovery, project teams share a common goal: to find an effective and selective clinical candidate as quickly as possible. That is certainly the goal at Galapagos, a commercial-stage European biotech specialized in the discovery and development of small molecule medicines with novel modes of action.
Traditionally, medicinal chemists have manually iterated on design, synthesizing and testing molecules in cycles that often took years. Computational methods can accelerate that process by enabling chemistry teams to explore vast stretches of chemical space and conduct experimental evaluation in silico, optimizing molecules for properties such as potency, selectivity, solubility and more.
Over the past several years at Galapagos, we have worked to incorporate computational workflows into our programs holistically – from project team collaboration and hypothesis tracking to computational-driven ideation and compound triage. We have applied this approach on programs in inflammatory, fibrotic and metabolic diseases amongst others.
Below are three of our core pillars we use to successfully integrate cutting-edge computational technology into modern drug discovery teams:
- Assemble a collaborative team that puts the modeler on strong ground: Collaboration is critical for any high-performance team. While this might seem obvious, it requires a delicate balance. Chemists and project leaders should never have sole ownership of a project – it’s vital to fully integrate all project members, including molecular modelers, into every aspect. Modelers should not be list-takers or service providers, relegated to support requests from the chemistry team. Instead, modelers should be seen as a crucial member of the core team, fully integrated into strategy discussions and proactively contribute to proposals.
- Inspire more 3D thinking by offering foundational training: Even with a robust computational platform, many project team members don’t necessarily know what to do with 3D modeling results. It’s vital to make time for advanced coaching of the broader chemistry team on fundamental modeling concepts – not just on how to use the technology, but how to interpret the results and utilize them to make decisions. Through efforts such as this, organizations can foster more 3D thinking among teams, educate team members to ask better questions of the computational tools, and ultimately build trust in the results.
- Foster model-enabled ideation to drive more quickly toward a development candidate. There are different breeds of ideation platforms – not all are alike. Finding a platform that fosters a “predict-first” mindset and enables the full project team to collaborate on a crowdsourced design is a critical starting point. At Galapagos, we found this platform in Schrödinger’s LiveDesign, an enterprise-scale informatics platform that enables teams to keep all their experimental and modeling data in one place, propose new ideas, iterate, and optimize. LiveDesign gives us a wide range of target-based and ligand-based modeling tools to utilize. It’s intuitive to use – the entire workflow for chemists requires just a few clicks. It consolidates best-of-breed scientific capabilities in a single interface. We are also actively leveraging Schrödinger for a new de novo design workflow we call “Project Cyclone,” an extension of LiveDesign for automated generation of molecules guided by machine learning predictions that use the latest project data.
Ultimately, integrating cutting-edge computational technology requires attention to all three of these strategies. It’s a tractable process. Science and computational power have made huge progress – but a lack of integration in workflows often hampers impact of approaches. By combining data and scientific workflows into a single interface with LiveDesign, we have made significant strides in empowering project teams to make model-driven chemistry decisions and achieve a higher collective potential to push programs forward efficiently.