We streamline protein design and development with MADITM, our proprietary deep learning algorithm, to predict high value protein products and enzymes with unprecedented speed.
Our in-house lab generates small quantities of the new predicted structures to validate their desired performance. Upon lab validation, we collaborate with our clients on next steps to move from pilot to industrial scale production.
We start a new program by defining a desired biochemical process. We input critical parameters, such as the structure of a target molecule or substrate the candidate protein needs to react on. For enzymes, we first search for an existing enzyme that performs a desired biocatalytic reaction that needs improved efficiency. MADI takes these inputs and works extremely fast testing millions of candidate protein structures, then updating and re-testing. After the first run of MADI, we generate 10 – 100 strong candidate protein structures and can begin an efficient lab screening to rank the best performing proteins. This is unlike competitive platforms that require costly screening of thousands of candidates relying less on prediction software to narrow the candidate pool.