BIOVIA Generative Therapeutics Design
Combine Human Ingenuity with AI for Faster Drug Discovery
Finding one viable lead drug candidate takes about five years and requires the synthesis of thousands of molecules. These numbers have not improved for decades. The time to change this is now!
Finding a lead drug candidate is challenging. It requires an optimization of many characteristics that together make up the candidate’s Target Product Profile (TPP). Traditional approaches are ineffective – improving one property often worsens another.
Transform Drug Discovery
BIOVIA Generative Therapeutics Design is a cloud-based integrated solution that optimizes drug design and discovery — potentially saving millions of research dollars:
- Combine advanced data science, machine learning, cheminformatics and structure-based modeling to explore chemical space
- Automate the virtual creation, testing and selection of novel small molecules
- Reduce the cost of physical testing.
Generative Therapeutics Design combines Virtual and Real (V+R) lead optimization to support the “active learning” innovation cycle:
- Virtual cycles: Explore chemical space by “learning” from real experiments. The system virtually screens and optimizes candidate compounds using a combination of Machine Learning models and structure-based modeling and simulation methods. Multi-objective optimization algorithms balance competing objectives, allowing teams to generate compounds that improve towards a TPP.
- Real cycles: Synthesize and test the most promising virtual compounds in the lab. Use the new data to improve predictive models and refine the exploration of chemical space. These V+R active learning cycles continue until you identify compounds that meet the TPP.
Critical Success Factors
- Modeling and Simulation methods are essential for improving the quality of compound suggestions. Promising approaches include pharmacophore scoring, molecular docking and free energy methods for predicting binding affinities. BIOVIA has over 20 years of experience in supporting computational scientists using these methods, which we are integrating into our generative design solutions.
- Synthetic Feasibility scoring helps scientists to design virtually generated compounds with fewer experimental cycles, while significantly reducing overall discovery costs.
- V+R Analysis and Decision-Making pairs the unique expertise and intuition of the chemist with AI predictive modeling to shape the decision about which compound to make next. All virtual and real information is then combined in an analytics environment that takes into account preferred chemistries and reactions and available resources/knowledge from ‘prior art’ to find the most promising compounds to pursue.
Cloud and Collaborative Platform – Molecular discovery is a team effort. Medicinal chemists work with biologists, computational scientists and data scientists. Many of these collaborations are now external. Common data models facilitate team collaboration and tech transfer to downstream activities. Hybrid Cloud, an integration between on-premises and cloud activities, ensures that the right people have access to the right information at the right time.
Generative Therapeutics Design transforms Drug Discovery to:
- Shorten lead optimization phase for drug candidates
- Reduce the number of synthesized compounds and experimental assays run per project
- Deploy an agile, secure and cloud-based solution with low total cost of ownership
- Expand the diversity of investigated molecules
- Consider existing intellectual property in the design process
- Increase success rate during early clinical stages.