Shiru CEO: AI-driven protein discovery revolutionizing biopharma, now moving into food

By Agroempresario.com

The protein discovery process in biopharma has been significantly disrupted by artificial intelligence in recent years, according to Dr. Jasmin Hume, the founder and CEO of Shiru. While the food industry has been slower to adopt AI tools, Hume believes the same disruption will soon happen in this sector as well. Shiru, a California-based startup, has developed a first-of-its-kind platform that uses AI and machine learning to help companies discover proteins for applications ranging from high-intensity sweeteners to binders and emulsifiers in the food and cosmetics industries.

Shiru’s platform, ProteinDiscovery.ai, enables users to search a vast database of millions of protein sequences, categorizing them by functional use, protein sequence, and successful expression — meaning how well the protein can be produced in microbes via precision fermentation. Hume shared this insight during the Future Food-Tech summit in San Francisco.

Unlike companies focused on designing synthetic proteins, Shiru explores proteins already existing in nature, aiming to discover the most functional proteins for applications in flavor, texture, and bioactivity. Shiru’s library consists of approximately 33 million sequences from plants, algae, and microbes. The company leverages bioinformatics and machine learning to organize this data based on each protein’s functional properties.

“We focus on the job the protein needs to perform in a food product, such as emulsifying or enhancing sweetness, and use that information to guide our search algorithms to identify proteins that can do the job,” explained Hume. Shiru can then produce samples of those proteins for partners who lack the capability to produce them in-house.

Additionally, Shiru has developed a system to predict how efficiently a given protein can be expressed in microbial hosts, ensuring it can be manufactured cost-effectively at scale. The company has also partnered with molecular farming startup GreenLab to express certain proteins in corn kernels, a scalable method for protein production.

“We’re applying tools that are not yet common in the food industry, such as large proteomic databases, biotechnology, AI, and machine learning, but we’re making them easy for companies to use,” said Hume. “Our goal is to make this a plug-and-play system where companies just need to know what their ingredient needs to do, and we’ll handle the rest.”