Foundational Model for Functional Molecules
The future of biology is designer and 310 is building the AI engine to power this reality. In this future, novel functional molecules with desired properties and function will be designed precisely, accurately, and almost instantly – delivering near limitless value to medicine, climate and industry.
Our multi-modal, multi-task approach will result in the best foundation, which will result in the best performance in downstream applications.
310’s AI engine is Large – spanning the broadest possible range of multi-context and multi-modal data, Active – self informing additional data to be collected, and Generative – performing end-to-end design against desired specifications
Highly multi-context & multi-modal
Model informs what data needs to be collected
Get exactly what you need
310 leverages the broadest possible range of multi-context and multi-modal data. This includes nearly 20 Terabytes of highly curated public and proprietary data. The data is continuously improved with addition of new high quality data sets.
Our model can generate novel (%SeqID < 80) protein sequences of variable length, of high quality (pLDDT > 80), and high structural match to a target (TM-Score > 0.8)