Key Challenges
- Client has a platform where researchers can access data regarding drug candidates, gene expressions, protein-protein interactions, and clinical records
- They wanted to integrate rapid analytics and machine learning capabilities into this application to enhance the drug discovery process
- They wanted to improve the probability of FDA approval to make drugs more affordable
Technologies
Solution
- We analyzed the existing architecture of the client’s application and re-designed it to support cognitive functionalities
- Data sets available involved millions of compounds, we cleansed and normalized data from internal and external sources (public libraries) for predicting therapeutic potential
- Designed and developed ML models for optimizing formulation conditions for proteins based on available in-house data and externally available stability data
- Implemented appropriate data governance and management principles for reusability of existing models
- Integrated with third-party modeling and visualization tools for enhanced/better drug discoveries using knowledge graphs
Benefits
- Helped them reduce research and development costs by 25%, while avoiding costly errors
- Improved regulatory compliance, and increased collaboration between teams and departments
- Enabled faster data-driven decisions to gain insight into business
