Demonstrate and verify the intellegent de novo drug discoverer
| Reference number | |
| Coordinator | ANYO Labs AB |
| Funding from Vinnova | SEK 1 459 000 |
| Project duration | May 2024 - November 2025 |
| Status | Completed |
| Venture | Banbrytande tekniklösningar |
| Call | Groundbreaking and scalable technology solutions in 2024 |
Important results from the project
Yes: we needed to verify the value and quality that our generative AI produces compared to traditional library screenings. The result was that 28 compounds were synthesized and tested, and although many were fluorescent, we did manage to set up an assay verifying that we could find valid hits. However, they were neither more potent nor achieved a higher hit rate than on-demand compound screenings.
Expected long term effects
This grant opportunity gave us the ability to truly see if generating novel chemical entities from step 1 in the discovery process is a worthwhile endeavour. We have learned to look into the planarity of the structures generated by our AI, while also being more aware of the high synthetic feasibility rate. We also know now when to implement generative AI vs virtual screenings for both partners and our internal pipeline we will be investing heavily in.
Approach and implementation
The activities were correct for the plan. A difference would have been to allocate the synthesized compounds to more than 1 project/target to understand the results depend of the pocket and target or the generative AI. Also, some in vitro testing after hit confirmation would have added real value to the end results, at least for the compounds that did bind to the desired target. Regardless, these are changes in hindsight and the results were still very rewarding to see.