Ersilia wiki
Merck Biopharma Speed Grant
Announcement of the Merck Biopharma Speed Grant - October 2021
EOSI has been awarded a Merck Biopharma Speed Grant for the project "A privacy-preserving machine learning toolbox for secure distribution and dissemination of drug discovery results protected by intellectual property"
Data-driven drug discovery research is on the rise, with novel AI/ML models being developed for all stages of the process. However, many AI/ML initiatives are hampered by undisclosure of experimental data from pharmaceutical companies, mainly due to intellectual property (IP) constraints. This trend applies to the earliest screening campaigns too, even if an overwhelming majority of results are considered to be negative and most compounds end up being attrited.
The benefit to the academic research community of accessing the wealth of knowledge produced in pharmaceutical companies is undeniable. We believe this data can be released using AI/ML models as a surrogate. Simply put, IP-protected datasets can be used to train AI/ML models that will then be donated to the community without having to disclose the identity of the training compounds. This would enable researchers working, for example, in malaria, to test their candidates against a highly proficient AI/ML model trained on proprietary data.
To this end, we aim to develop an encryption toolbox that will protect the identity of the underlying compounds and couple it to our Automated ML pipeline. The project will run from October 2021 to July 2022
Gemma Turon (CEO) - A privacy-preserving tool is a key step towards accessing data that is currently kept in the drawer, data that can bring us closer to finding new drugs for the most pressing neglected diseases.
Last modified 1mo ago
Copy link