Some proteins can perform useful therapeutic functions but remain difficult to store, transport or manufacture. Scala Biodesign is building a computational service to suggest changes that could make these molecules more practical, using protein structure predictions alongside clinical data and observations of naturally occurring proteins.
The company emerged from research at the Weizmann Institute of Science in Tel Aviv and raised $5.5 million in seed funding. Its approach aims to replace broad trial and error with a smaller number of targeted protein designs for researchers to evaluate.
Why protein engineering can take so many attempts
A promising protein may have a weakness that limits its use. It might break down at room temperature or be vulnerable to the chemical environment inside the body. A change to its molecular sequence could help, but researchers first need to work out which change is likely to improve the desired property without undermining the protein’s function.
That search can become enormous. The article describes a small protein made of 100 amino acids, with 20 possible options at each position. Trying every possible combination is impractical, and random attempts to find an improvement can take a long time, fail, and cost millions.
Scala CEO and co-founder Ravit Netzer said protein development at even large companies is largely trial and error, often using random mutagenesis. With more knowledge of protein structures, she said, randomly changing parts of a molecule is no longer a sensible default.
Using data to narrow the search
Scala combines protein structure prediction with clinical data and observations of naturally occurring proteins. The company says this helps its system identify sequence changes that may support a specific goal, such as improving stability, increasing a protein’s effect or making it easier to manufacture.
The process is computational and does not involve a wet lab. Scala provides a small set of candidate sequences that it considers high confidence, with the aim that at least one will move the protein in the desired direction. The approach still produces options rather than a guaranteed single answer.
Netzer compared the challenge to replacing a random word in a paragraph and hoping the result improves it. The company’s goal is to make changes guided by information about the molecule, instead of searching blindly through combinations.
A malaria vaccine example
One lab was working with a naturally occurring protein that functions as a malaria vaccine. The protein was sensitive to temperature, creating a concern that it might not withstand transport or storage.
Scala CTO and co-founder Adi Goldenzweig said the lab supplied one input and received three outputs. The team selected the best candidate, which is now in clinical trials. He said the ideal would be to provide one option with complete confidence, but the company is not there yet. In comparison, researchers may otherwise work through tens of thousands of possibilities.
The proposed changes can also extend beyond a single amino acid. Goldenzweig said that, for larger proteins, Scala may swap in dozens at a time, including more than 50 mutations in one shot. That scale of simultaneous change is part of the company’s effort to explore options that conventional trial and error may not readily reach.
Building partnerships to prove the approach
Netzer said Scala has validated protein design across diverse applications, including antibodies and enzymes. The company wants to show that meaningful improvements can be designed at scale, beyond work limited to an individual research project.
Scala is working with unnamed pharmaceutical companies and labs. It is keeping its licensing and business model flexible, while prioritizing the service and proof of its technology over building its own biological intellectual property. The founders have not ruled out developing that kind of IP in the future.
The $5.5 million seed round, led by TLV partners, was the company’s first. After emerging from stealth, Scala plans to pursue more partnerships and studies. Its broader ambition is to make protein engineering easier to access, while its immediate task is to demonstrate that its suggestions can help researchers improve useful molecules.