A voluntary participation initiative to enable improved predictive insights and better, faster outcomes in drug discovery
SynSight is an opt-in consortium of clients who consent to anonymized usage of their data towards predictive models with higher reliability and broader applicability, leading to better decision-making, timelines, and outcomes on their projects. It provides a common platform for members to benefit from the learnings on larger volume, standardized data while protecting proprietary information.
We envision SynSight to facilitate quality predictive insights that enable informed decisions for its participating members. Being part of the consortium allows access to learnings from broader data that have better accuracy and applicability. Members can leverage these insights to improve the speed and quality of their projects, leading to shorter design-make-test-analyze (DMTA) cycle times and better outcomes.
Ensuring data confidentiality
SynSight ensures data confidentiality by applying solutions such as data anonymization and federated learning. Anonymization embeds molecular data into fingerprints or features, which are used to train the models. The solution makes reconstruction of the parent data practically impossible.
The parent data of the members is not aggregated. Only anonymized data is aggregated to train the models. The models, thus, do not retain any molecular information from the parent data.
Where possible, federated learning is used to add an additional layer of data confidentiality.
To know more about our data consortium, download the brochure.
