This undisclosed RNA manufacturer needed to move beyond trial-and-error with LNP design. Synthesising and testing molecules was adding years and runaway costs to development.
Abzu created a real-time LNP designer to predict apparent pKa value.
Your partner for precise and exceptional results — a LNP design application in just two months.
We supported two Computational Chemists at an undisclosed RNA CDMO to develop a LNP designer that predicted apparent pKa value.
Reengineering features required collaborative research. We were able to cross validate the client’s features with new feature descriptors that we discovered in research to identify the best structure-type descriptors for LNP design.
We trained models on the client’s data and selected the best-performing model. This was validated against a set of molecules that had not been previously seen by the model or the client’s team.
Abzu revealed the mechanisms behind apparent pKA value.
Chemical understanding, enhanced predictions, and a practical application.
By modifying properties to see — and understand — the impact of changes, our client was able to screen in silico in a week what would have taken them years to synthesize and test.
Less data, more discovery.
Make the most out of every piece of information.
Contrary to the data-heavy demands of black-box AI, our technology thrives on minimal data. This efficiency not only accelerates the discovery process, but also reduces your barrier to entry to develop new, innovative drugs.
