University of Washington spinout A-Alpha Bio snags $20M for protein-discovery platform – GeekWire
A-Alpha Bio will build lab space in downtown Seattle and beef up its machine learning team with $20 million in new venture funding for its operation to identify therapeutic proteins, the company announced Thursday.
Seattle’s Madrona Venture Group, a tech-focused firm branching into the life sciences, led the Series A round.
The funding propels the company from a grant-funded University of Washington spinout with $2.8 million in seed funding to a full-fledged biotech supporting the development of potential treatments for COVID-19 and other diseases.
And the more proteins the company analyzes, the better it gets at it. A-Alpha Bio uses information from its datasets to train its algorithms to predict favorable protein interactions.
“Over time we’re accumulating these enormous repositories of protein interaction data that we can also use with computational tools like machine learning to better understand and engineer protein interactions,” said CEO David Younger in an interview with GeekWire.
Younger founded the company in 2017 with CTO Randolph Lopez using the technology they helped to develop while graduate researchers at the UW’s Institute for Protein Design.
The drug development platform company combines computational tools with yeast experiments to identify potentially therapeutic proteins. It partners with drug companies to help them find the best agents to test on a range of conditions.
The technology relies on pools of single-celled yeast engineered to express various proteins or protein fragments. When two proteins interact, the yeast fuse and the interacting proteins are tallied. The method can identify protein interactions such as an antibody sticking tightly to a viral protein.
A-Alpha Bio hits the sweet spot of high accuracy and high volume of sampling to measure protein interactions, said Younger.
“Historically there’s been this really dramatic trade off, where either you’re getting high quality or you’re getting high quantity, and it’s very hard to maximize both,” he said.
A-Alpha Bio has partnerships with three large pharma companies and is aiming for more. It is also partnering with the Bill & Melinda Gates Foundation, Twist Biopharma and Seattle-based Lumen Biosciences to help develop COVID-19 antibodies, among other projects. Lumen grows therapeutic proteins in the algae spirulina and recently announced plans to build a new production plant in a former Seattle bakery.
A-Alpha Bio’s high accuracy and throughput give it an edge over competing platform technology companies, said Younger. Other companies in the space include Adimab and Distributed Bio.
Younger and his colleagues recently showcased their technology in a study identifying antibodies to variants of the COVID-19 virus. Such antibodies could potentially be used as therapies to treat the disease.
The researchers started with yeast expressing a variety of antibodies as well as proteins from different virus variants. “And we throw those all into one pot, map all of the interactions and then determine which antibodies are going to have the right profile,” said Younger. The paper appears in the preprint server BioRxiv and has yet to be peer reviewed.
In addition to identifying antibodies, A-Alpha Bio is focused on finding targets for “molecular glues.” Such “glues” connect a target protein with another type of protein that tells the cell to eliminate the target.
“It’s essentially a way to target particular proteins for degradation,” said Younger. And it’s also the basis for the chemotherapy drug Revlimid. Revlimid has pulled in billions in revenue for Celgene, a Bristol Myers Squibb company, and spawned intense interest in the area of molecular “glues.”
A-Alpha Bio is also leveraging computational methods developed by the IPD for their tool predicting how proteins fold. IPD’s tool, called Rose TTAFold, and a similar one developed by DeepMind recently blew away the life sciences community with their speed and accuracy in predicting protein folding.
Those tools were made possible by existing massive datasets on the folding of known proteins, whereas the training sets for predicting protein-protein interactions are still limited. But A-Alpha Bio sees that changing as it builds up its databases.
A-Alpha Bio can measure millions of protein interactions a single experiments, said Younger, “And so with this volume of data we can start to train machine learning models that essentially learn the rules of protein-protein binding.”
“Our vision is to get to a point where we can essentially do in silico antibody antibody prediction for a given never-before-seen target,” he added.
The combination of lab science and computational approaches is an emerging field of investment for Madrona, managing director Matt McIlwain previously told GeekWire.
The decades-old venture firm has traditionally invested in tech companies but is increasingly backing companies at the “intersection of innovation.” These companies at the interface between biology and computer and data science also include Nautilus Biotechnology, TwinStrand Biosciences, Ozette, and Terray Therapeutics.
“As machine learning, computing power, and wet lab automation advances converge, we expect to see even more companies approaching research problems in paradigm shifting ways,” McIlwain and Madrona partner Chris Picardo said in a blog post about the new funding.
A-Alpha Bio currently has 13 employees and expects to expand to about 50 within the next two years while building out its machine learning team. Prior to the pandemic, “We had a pretty strong conviction around keeping the team as local as possible,” said Younger.
But like many operations, A-Alpha Bio has learned to operate effectively remotely and is open to hiring off-site. Said Younger: “We’ve learned recently that we do not, under any circumstances, want to sacrifice talent.”
The company is also converting existing office space in the Belltown neighborhood for its new lab for experiments in yeast, a creature that does not require a lot of upkeep.
“One of the really great things about working with yeast as a model organism is that our needs from an infrastructure perspective are very simple,” said Younger, “Instead of having to fight tooth and nail with other biotech companies for that 1% of vacant lab space in Seattle, we can move into office space and do a full conversion into a lab that meets all of our needs.”
The company is the latest IPD spinout to snag new funding after the recent IPO of Icosavax, a company developing vaccines for COVID-19 and other diseases.
Previous seed funding was led by OS Fund, and the company has also received grant support from the Gates Foundation and the U.S. National Science Foundation. In addition to Madrona, participants in the Series A funding round include Perceptive Xontogeny Venture Fund and Lux Capital.
Xontogeny’s Ben Askew and McIlwain, who is also on the board of trustees for Fred Hutchinson Cancer Research Center, have joined the company’s board of advisors.