In the modern world, artificial intelligence (AI) and deep learning have helped to accelerate important advancements in technology across a wide range of industries, with the biotechnology sector being a prime example of such transformative applications. In Hamilton, Ontario, one biotech company is harnessing the power of deep learning and AI to accelerate the discovery of new pharmaceuticals from microorganisms found in nature and even the human body.
Adapsyn Bioscience, a spinout company from McMaster University, mines the landscape of highly-evolved small molecules from microorganisms to uncover a diverse portfolio of novel drugs to treat a wide spectrum of diseases. Adapsyn was incorporated in 2016, and soon after its inception drew the attention of pharmaceutical giant Pfizer, entering into a collaboration to accelerate their drug discovery efforts. The company now has 12 full-time employees, in addition to several others who are on a part-time or contract basis.
“I’m one of those Canadians that was in the U.S. for a long time, and then decided to come back to Canada,” says Adapsyn founder, president & CSO, Nathan Magarvey. “I came back with aspirations of setting up a company to engage in research here rather than the U.S. Adapsyn was a growing body of work that was established from my research lab, and the lab was increasingly taking on industrial work. It started getting a bit too much for an academic lab, so that’s when I said it’s probably time to spin out and make it grow. Building a start-up is an experience unto itself, and something I always wanted to do.”
Subsequent to the research deal formalized with Pfizer, the company raised venture capital from Genesys Capital in Toronto, a mainstay in the Canadian biotech sector, and Pfizer Ventures out of the U.S. The company strengthened its board of directors by adding representatives from these firms and additional industry experts.
“The microbiome is increasingly a focus for the company,” says Magarvey. “We’ve been engaged in the microbes of the planet and sifting through those genomes to identify molecules. What Adapsyn does better than anyone is translating genes to proteins to small molecules. Those evolved small molecules are highly valuable against a whole spectrum of human therapeutic drug targets and can be particularly useful as anti-infective and immunomodulatory agents. Adapsyn is able to translate the gene sequence directly to the small molecule structure, and can, in turn, determine which structures are new and which ones are old, all from the genome sequences that we have. We can then predict from that data which ones do new things and which ones do things we have seen before from known compounds.”
This is really at the heart of the machine learning and AI that the company uses to inform its wet lab activities – namely, purifying and testing novel molecular entities for therapeutic potential. The company has created a data set of all-natural products discovered to date, and uses this to relate predicted novel compounds to past discoveries, thereby allowing the company to infer potential uses of newly discovered natural products.
“For a long time, we would focus on microbes and plant sources. But the challenge with that, by modern day standards you are a prisoner of the past. All the molecules that were identified previously, if you use old methodologies, you end up discovering the same molecules that you previously found. With our technology, we never find the compound we found before, and given that around 1 in 200 molecules become clinical candidates, we can really ramp up programs to push those novel entities into therapeutic models,” says Magarvey.
Adapsyn’s technology allows them survey genomes of microorganisms, numbering over a hundred thousand to date, and rapidly pinpoint novel chemistries. With this pace of discovery, the company has seen a flourish of interest from many pharmaceutical companies. While some companies have enormous depositories of biological material from which to screen, they are inadequately-equipped to rapidly assess their chemical novelty. Adapsyn has proven they possess the technological platforms necessary to unlock this potential.
“This is high science. This is a merger of chemistry, microbiology, genomics, and computer science,” says Magarvey. “Canada does well in certain elements of each field. The Canadian research community is also recognized for its creativity in the connection disparate research fields. Adapsyn is an exemplary showcase of this creativity and scientific translation. Striving to address the unmet medical need is where we can make a transformative impact on Canada and globally.”
“It’s fascinating for us to appreciate now all the medicines that are left in nature that are awaiting discovery to address more complex diseases,” comments Magarvey. “We don’t appreciate this in society, but a lot of our medicines – not just your holistic things – but medicines that we take are actually derived from living creatures on the planet. With our new technology, we can sort through all that information so rapidly, that now we see all these new opportunities.”
Drug discovery can be a long, complex and expensive process, but with Adapsyn’s technology, it could take less time from bench to the bedside.
“We’re getting better and better at discerning the molecular targets of those predicted molecules by using deep learning,” Magarvey says. “That’s really important, because obviously you have to be able to quickly identify what the molecular targets would be of molecules and their value as therapeutic entities. So, we’re using cell-based, high-content screening, in combination with our predictions, and deep learning is having a big role in elucidating the mechanisms of action of these novel compounds.”
Adapsyn Bioscience is altering the drug discovery landscape, combining genomic and metabolomic data with AI and machine learning. Their unique approach will help fuel the future of natural products research, ultimately leading to the development of new pharmaceuticals that will impact global healthcare and address the treatment of diseases that currently lack effective therapies.