Algorae Pharmaceuticals advances artificial intelligence drug prediction platform

Go to Imelda Cotton author's page
By Imelda Cotton - 
Algorae Pharmaceutical ASX 1AI AI supercomputer GADI

Algorae Pharmaceuticals (ASX: 1AI) provided an update today on how it is advancing artificial intelligence (AI) biopharmaceutical prediction platform.

The Algorae Operating System (or AlgoraeOS) platform is designed to predict synergistic combination drug targets using machine learning, deep learning and artificial neural network algorithms within a “vast” database containing medical and scientific information curated for the purpose of AI drug discovery.

It is being developed in collaboration with AI experts from the University of New South Wales and version 1 is expected to be launched in Q3 (July-September) this year.

Algorae intends to generate value from the new platform by using AI to generate fixed-dose combination drug targets that will either expand Algorae’s internal Food and Drug Administration therapeutic pipeline or be licenced to third parties.

This is a revolutionary new type of business model, often known as “techbio”, one which is gaining significant attention in the United States.

Some of the more advanced techbio companies that have demonstrated their ability to generate drug candidates using AI include Recursion Pharmaceuticals and Relay Therapeutics, both of which have market values exceeding US$1b.

Key early drivers of their value have been either drug development partnerships or the discovery of drug candidates they are taking to the clinics on their own.

New supercomputer

In today’s announcement, Algorae detailed that it will now be run on ‘Gadi’, the most powerful supercomputer in the Southern Hemisphere.

Gadi is operated by National Computational Infrastructure Australia and has previously been used for climate modelling and natural disaster prediction.

Gadi – which means ‘to search for’ in the language of the Ngunnawal people – contains more than 250,000 central processing unit cores, 930 Terabytes of memory and 640 Nvidia graphics processing units.

The supercomputer has peak operational capacity exceeding 10 petaflops, or 10 quadrillion floating-point operations per second, that the company says will provide significant computational leverage to facilitate the AI interrogation of the scientific and medical information within the AlgoraeOS database once completed.

The build-out advances

Data acquisition and customisation is currently underway across the company’s four pillars of information: drug data, cellular data, chemistry and biological data.

The data will cover scientific and medical fields including drug chemical structure, drug target interactions, gene expression and drug perturbation.

Algorae stated today that data modules will be added to its database progressively and the AI platform will be refined in future versions over a period of at least three years to improve its predictive capabilities.

Apart from project leader and AI expert associate professor Fatemeh Vafaee, the company has previously appointed three full-time development staff to the AIgoraeOS research team.

Algorae intends to add a fourth full-time member to the AI team and is currently interviewing candidates with sophisticated pharmacology experience.

Funding for the team is partly provided by a 3-year Commonwealth Scientific and Industrial Research Organisation grant that Algorae was successfully able to procure in October 2023.