DeepCure, founded by researchers from the MIT Media Lab, develops disruptive Artificial Intelligence to revolutionize pharmaceutical development. The pharma industry is undoubtedly one of the world's largest, with worldwide revenues topping one Trillion US dollars. However, its effectiveness is in question, as the exponentially increasing cost of drug discovery and development has resulted in fewer new drugs reaching patients at a much higher price. Development of a single new drug today cost over 2.5 Billion US dollars, an amount that went up from a 100 Million in just 50 years. This trend is a critical barrier to the development of new and improved drugs, and is making the existing medicines unaffordable for most of the global population. At the origin of this problem, lies our inability to exhaustively search the astronomical drug-like molecules space for an efficacious drug to a disease-related biological target. To support innovation and reverse the decaying trend, we need a paradigm shift in the form of efficient exploration methods.
The inefficiency of small-molecule drug development is also promoting the rise of biological therapeutics (biologics). Biologics are a tremendous pharmaceutical innovation, which improves many lives around the globe and provides cures to diseases that were untreatable just a few years ago. However, biologics also form serious accessibility challenges to the healthcare community. In most cases, biologic treatment requires frequent administration by a trained health practitioner, sometimes even multiple times per day. Moreover, an unbroken cold chain is usually mandatory for the transportation of biologic drugs. These requirements are unrealistic for the majority of the world population and, combined with the high cost of biologic treatment, decrease treatment accessibility.
In DeepCure we create innovative algorithms that combine the power of Deep Learning, Cloud Computing, and our proprietary database of one trillion chemistries to improve the breadth and accuracy of pre-clinical drug development. Our unique approach pushes the efficiency boundaries of small-molecule drug development by searching a molecular space that is a million times larger than currently possible, and doing so while minimizing the laborious, expensive, and error-prone lab work. DeepCure's platform finds the optimal target-drug interactions, while, at the same time, accounting for the dynamic nature of the target, optimizing the required physical properties of the drug, and minimizing possible off-target effects. Moreover, our algorithms guide the human chemists in the lead optimization step to ensure better small molecules get to the clinic. Overall, DeepCure's toolkit saves up-to 80% of the pre-clinical time and cost, compared to industry standards. More importantly, our algorithms have the potential to find highly effective drugs which do not exist in any existing drug library and are extremely unlikely to be discovered by any existing conventional drug discovery pipeline, including small-molecule alternatives to inaccessible biological therapeutics.
As a primary technological validation, our algorithms rediscovered one of the top-selling small-molecule drugs in the market (an RNA polymerase inhibitor for the treatment of Hepatitis C with $14B Billion in annual sales) in just a few weeks and suggested promising candidates for a highly desirable oncology target.