Allozyme's ingenious method of rapidly testing millions of bio-based chemical reactions has proven not only to be a useful service, but also the basis of a unique and valuable data set. Where there are datasets, there is AI, and where there is AI, there are investors. The company just raised $15 million in Series A to grow the business from a helpful service to a world-class resource.
We first covered this biotech startup in 2021, when the company was taking its first steps. “At the time, we were less than five people and our first lab was 1,000 square feet,” recalls CEO and founder Peyman Salehian.
The company has expanded to 32 employees in the U.S., Europe and Singapore, and is using 15 times more lab space to accelerate its already dramatically faster enzyme screening technology.
The company's core technology has remained unchanged since 2021, and you can read its detailed description in the original article. But the bottom line is that enzymes, chains of amino acids that play specific roles in living systems, have been quite difficult to discover and invent. That's because the number of variations is huge. Molecules can be hundreds of acids long, with 20 acids to choose from at each position, and every permutation can have a completely different effect. Instant access to billions of possibilities.
Using traditional methods, these mutations can be tested at a rate of hundreds per day in reasonable laboratory space, but with Allozymes, by packing the enzyme into tiny droplets and letting the enzyme pass through. , using a method that can test millions of enzymes a day. Special microfluidic system. You can think of it like a conveyor belt with a camera on top that scans each item you zoom in on and automatically sorts it into different bins.
Droplets containing enzyme variants are evaluated and redirected to the microfluidic system if necessary.Image credit: Allozyme
These enzymes can be anything needed in biotechnology or the chemical industry. When a raw material needs to be converted into a specific desired molecule or vice versa, or when it needs to perform many other fundamental processes, enzymes do it. Finding something cheap and effective is rarely easy, and until recently the industry was testing about 1 million possibilities a year, but Allozyme has increased that number by more than 1,000 times. The goal is 7 billion variants by 2024.
“[In 2021] We just built the machine and now it's working very well, screening up to 20 million enzyme variants per day,” Salehian said.
The process has already attracted customers in many industries, and while Allozymes can't disclose some due to non-disclosure agreements, others are documented in case studies.
Phytoene is an enzyme naturally found in tomatoes, typically harvested in small quantities from the skins of millions of tomatoes. Allozyme has discovered a route to producing the same chemical in a bioreactor using 99% less water (and possibly space). Bisabolol is another useful chemical found naturally in the candeia tree, a plant native to the Amazon that has been driven to endangered species status. Now, using a bioreactor and the company's enzymatic pathway, bioidentical bisabolol can be produced in any quantity. Fiber from plants and fruits such as bananas can be converted into substances called “soluble sweet fibers'' that can replace other sugars and sweeteners. Allozyme Inc. won him a $1 million grant to accelerate this not-so-easy process. Salehian made cookies and bubble tea and reported the results.
I asked about the possibility of microplastic-degrading enzymes. Enzymes have been the subject of much research and have even appeared in Allozyme's promotional materials. Salehian said it is possible, but not economically viable at this time with the current business model. Basically, customers have to come to the company and say, “I want to pay to develop this.'' But it's on their radar and they may soon tackle plastic recycling and handling.
So far, this all more or less applies to the company's original business model, which amounted to enzyme-as-a-service optimization. But the roadmap includes expanding to more ground-up work, such as finding molecules that fit their needs rather than improving existing processes.
Allozymes' enzyme preparation service, which will be called SingZyme (as it does for single enzymes), will continue to be an entry-level option for people who want to “do this 100 times faster or cheaper.” This will satisfy the use case. A more expansive service called MultiZyme takes a higher-level approach to discovering or purifying multiple enzymes to satisfy a more general “I need something that does this.”
However, the billions of data points collected as part of these services will remain intellectual property and will constitute “the world's largest enzyme data library,” Salehian said.
CEO Peyman Salehian and CTO Akbar Vahidi, co-founders of Allozymes.Image credit: Allozyme
“When you give AlphaFold a structure, you know how it will fold, but you don't know what will happen when it's combined with another chemical,” Salehian said, adding that, of course, the industry is concerned. All that is happening is that reaction. “The data we have is so small and fragmented that there is no machine learning model in the world that can tell us exactly what to do. 300 samples a day for 20 years. ”This number can easily be exceeded in one day with the Allozyme machine.
Salehian said the company is actively developing machine learning models based on the data it has and also testing them based on known results.
“When we fed the data to the machine learning model, it came back with suggestions for new molecules that we had already tested,” he said, calling this a promising initial validation of the approach.
This idea is by no means unprecedented. We have covered numerous companies and research projects that have found that machine learning models can be extremely useful in classifying large datasets, providing additional confidence even when the results cannot be translated into real processes. Ta.
The $15 million A round includes new investors Seventure Partners, NUS Technology Holdings, Thia Ventures, and ID Capital, with reinvestment from Xora Innovation, SOSV, Entrepreneur First, and Transpose Platform.
Salehian said the company is in good financial shape and has plenty of time and money to achieve its goals. The exception is that it may raise a small amount of capital later this year to fund its expansion into pharmaceuticals and the opening of a U.S. office.