Could this Aussie AI start-up phase out lab testing on animals?
A Brisbane company has partnered with tech giant Google to solve a dark secret plaguing the drug development industry.
A Brisbane AI start-up company has partnered with tech giant Google in hopes of solving a dark secret plaguing the drug development industry.
Queensland-based biotech company Gelomics is using AI to help grow human tissues that could one day replace the mice and cell cultures that currently dominate the drug development industry.
It says its system more accurately predicts how real patients will respond, potentially saving lives, money and hundreds of millions of animals.
The company is aiming to improve drug development by using AI to assist pharmaceutical companies in growing human tissues that “look, behave, and react to drug treatment just like real human organs”.
Co-founder and CEO of Gelomics, Dr Christoph Meinert, said that “drug discovery and drug development are still heavily reliant on cell and animal experimentation”.
“Last year alone, close to 200 million animals were euthanised for research purposes, and around a hundred million or so cell culture experiments were performed,” he said.
“Yet, despite this massive effort, over 90 per cent of drug candidates that pass successfully through these testings will fail when they’re tested in human clinical trials.”
“The failure of every single drug candidate causes losses that exceed US$1.3 billion ($2 billion). And the total loss in the industry exceeds US$100 billion ($153 billion) every single year.”
Dr Meinert said biologists have a running joke that “every disease is curable in mice”.
“The problem is that models are relying on animal physiology, which has fundamental differences in the biology of animals compared to humans; they process drugs differently, the toxicity profiles are different, and also the efficacy testing is completely different,” he said.
“So the correlation between testing results you receive from animal trials and animal experiments to actual human results is very, very little.”
Dr Meinert says in oncology or cancer treatment, the correlation was less than five per cent, and in some other indications, such as Alzheimer’s, it was less than one per cent.
“That means that less than one per cent of drugs that are successfully tested in animal experiments achieve efficacy or a low level of toxicity in human trials. It’s a shocking number.”
Gelomics has just partnered with Google to complete the 10-week accelerator program through which they were able to build a fully operational end-to-end program.
By the end of the program, which partners AI start-ups with Google resources and mentors, Gelomics recorded 90 per cent predictive accuracy in the key parameters, and 60 per cent faster data analysis.
The program functions over four models: literature review, design, execution and analytics.
“One of the slowest parts of the entire process is researching what’s already been done here,” said Gelomics CTO Pawal Mieszczanek.
“The platform does this instantly for you. A researcher can simply enter a question and the system automatically summarises the latest scientific literature. Our operating papers give insights to the scientists before designing their experiments.
“In the design and execution module, the module guides the user through its entire process from tissue selection all the way to the parameter optimisation. The user can simply select the tissue type they want to grow, and our platform will optimise all the processing parameters for them in the background.
“Finally, their analytics module here, the experimental data can be analysed all in one place. Researchers can simply select the analysis they want to run, and the system generates clean visual insights for them.
Genetics PhD student at the University of Otago, Emily Morice, through the course of her research on how genetic differences contribute to inflammatory diseases, works with cell culture models extensively, and said Gelomics is operating in an exciting and scientifically meaningful space.
“Their 3D cell culture systems, biomaterials, and AI integration address long-standing problems in the field, particularly in reproducibility issues, limited throughput, drug discovery, toxicity screening, and early mechanistic work,” she said.
“Gelomics can make real gains in making 3D systems measurable, standardised, and reproducible. With their products that can make industry-wide standards with consistent hydrogel compositions (in which they grow the 3D cultures), organoid protocols, and validated analysis and quality control pipelines.”
Ms Morice added that drug failure is complex and multifactorial, and not just due to differences in species physiology.
“Failures often result from things like early-stage design (what the hypothesis is based on), use of models, and then during trials, things like dosing, pharmacokinetics, differences between patients (men vs women, under-represented groups and minorities), and then commercial issues like competition and market viability.
“It’s not that only one per cent of animal data ever predicts human outcomes; it represents a variety of factors over drug development as a whole.”
Despite Gelomics’ promise, Ms Morice said the bottom line wasn’t as simple as ‘animal vs non-animal’ testing.
“It’s using animal models selectively where our current technology lacks, and bolstering those studies with human-relevant models, each will contribute to our understanding of drug development. Companies like Gelomics, developing 3D systems, are important, but it’s important we use these findings in combination with other advances and not as a total solution.”
More Coverage
Originally published as Could this Aussie AI start-up phase out lab testing on animals?
