Algorae Pharmaceuticals launches AI-powered platform to generate novel drug candidates

Algorae Pharmaceuticals 1AI has launched Version 1.0 of its biopharmaceutical platform AlgoraeOS after construction at the University of New South Wales, now ranked the 19th best university in the world. The AlgoraeOS platform aggregates medical and scientific data on over 5,000 known drugs and molecules in over 150 human cell types. The platform uses trained artificial intelligence (AI) systems to predict fixed-dose combination (FDC) drugs that the company says it will either take to clinical trials itself or do so in collaboration with partners, potentially big pharma. The value of FDCs FDCs are pharmaceutical formulations that contain two or more active ingredients in a single dose formulation. When successfully discovered, FDCs can exhibit enhanced efficacy and / or lower side effects. “We’re using known scientific and medical data within AlgoraeOS to predict the discovery of FDCs that hold potential to improve medicines already in the healthcare market,” Algorae director Brad Dilkes said. Lower-cost strategy Algorae’s strategy is slated to develop new therapies without incurring the high cost and time burden of discovering and developing new molecular entities. “By focusing on existing drugs, formulated in novel combinations as predicted by AlgoraeOS, we have the opportunity to bypass the extensive and costly discovery phase of the R&D cycle for de novo (from scratch) drugs,” Mr Dilkes said. “Furthermore, developers of FDCs are generally entitled to accelerated development pathways with regulatory agencies such as the FDA”. If an FDC is novel, inventive and useful, it has the potential to be patented at the exclusion of competitors. AlgoraeOS has been trained to predict synergistic mechanism of action between any two drugs to form part of its inventive strategy to generate new intellectual property. On that point, Algorae has reported correlation coefficients of 0.91 – 0.98 “actual” vs “predicted” on various scientific measurements of synergy resulting from their training data in Version 1.0 of AlgoraeOS. The ‘patent cliff’ New intellectual property profiles over FDCs have the potential to be valuable to the biotech industry as big pharma races to protect revenues due to the slowing of blockbuster discoveries and impending patent expirations. About 190 FDA registered drugs are expected to lose patent exclusivity by 2030, including 69 ‘blockbusters’ (those with in excess of US$1b in annual sales). When a drug’s patent expires, the company that developed it loses its exclusive rights to manufacture and sell the drug. Gadi supercomputer AlgoraeOS is hosted by the ‘Gadi’ supercomputer operated by National Computational Infrastructure Australia and previously used for the likes of climate modelling and natural disaster prediction. Gadi contains more than 250,000 central processing unit cores, 930 terabytes of memory and 640 Nvidia graphics processing units. It has peak operational capacity of over 10 petaFLOPS, which exceeds the computational capacity of many other AI-enabled biopharmaceutical prediction platforms. Algorae has reported that its next steps include lab testing of AI-predicted FDCs, intellectual property assessments and partner evaluations.