If you are working on a GPCR project (or any other project lacking a protein structure) and cannot follow the traditional structure-based design approach, mining a virtual chemical space will be a useful approach to generate inspiring new chemistries as starting points.
Boehringer-Ingelheim used this approach to propel a GPR119 project. They started out with 4 structurally different literature compounds as queries to search their in-house virtual, BI-CLAIM to generate 10,000 FTrees hits. These were post-processed by a 3D shape filter based on one of the more rigid query compounds. Specific cores of the molecules were selected by the chemistry team as "activity anchors" and focused libraries around these designed and synthesized.
This approach lead to two new GPR119-agonist hit series, one of which subsequently resulted in a new lead class. Read more details in the original publication:
KnowledgeSpace™ is freely available chemical space covering hundreds of billions of virtual products. It can be searched out of the box, but being based on literature reaction protocols and commercially available reagents its IP-value is low. However, since we provide not only the space but also the reactions and the entire workflow to generate it, you may use this as a blueprint and design your own space using KnowledgeSpace™ as a jump-start. You can easily raise the IP-value by augmenting the space with your own proprietary reaction protocols and also by using your in-house building block collection.
Evotec did that. They used the KnowledgeSpace™ as a foundation and added their collection of 36 Mio in-house compounds (EVOsource) and several in-house reactions. The resulting virtual chemical space EVOspace comprises hundreds of billions of virtual products.
Many companies have a unique collection of in-house chemistries and building blocks, and therefore want to set up their own space, based on those in-house chemistries and building blocks. This will also create the highest level of IP-value. Pfizer did that and built PGVL. They turned 358 combinatorial libraries based in validated reaction protocols into a virtual chemical space, resulting in 3 trillion chemically accessible compounds. Read more in the original publication:
Don't see your case reflected here, or have a drug discovery problem you need a solution for? We can help
Simply contact our service team and ask for a free assessment.