Latent Lab launches web-based AI models to democratize protein design

Approximately six months after leaving stealth $50 million In fundraising, Latent Labs has released a web-based AI model for programming biology.

According to Simon Kohl, CEO of Latent Labs and a scientist who previously co-led with Deepmind’s Alphafold protein design team, the Latent Labs model “achieved cutting edge with a variety of metrics” when testing proteins generated in physical labs. The cutting edge Sota is a term commonly used in the AI field, representing the best performance ever in the industry on a particular task.

“There are calculation methods that evaluate the good design,” he told TechCrunch, adding that testing in the lab will result in a high percentage of proteins that the model creates viable.

The company’s basic biology model, known as Latentx, allows academic institutions, biotechnology startups and pharmaceutical companies to design new proteins directly in their browsers using natural language.

Latentx goes beyond what is found in nature and creates entirely new molecular designs such as nanobodies and antibodies with precise atomic structures. This approach helps develop new therapeutics with much faster rares.

According to Kohl, this ability to design completely new proteins distinguishes Latentx and Alphafold.

“Alphafolds are models for protein structure prediction. Therefore, they can visualize existing structures, but not produce new proteins,” he said.

In contrast to AI-driven drug discovery companies Xaira, Re-return Or deep mind spinout Isomorphic LabLatent Labs’ business model focusing on developing its own medicines includes licensing the model for use by external organizations.

“Not every company is in a position to build its own AI models, have its own AI infrastructure and have its own AI team,” Cole said.

Although Latentx is free to use, Kohl said the company intends to charge for the advanced features and features that were eventually introduced.

Other companies that provide open source AI basic models for drug discovery include Chai Discovery and Evolution Scale.

The Latent Lab is supported by Radical Ventures, Sofinnova Partners, Google’s chief scientist Jeff Dean, Anthropic CEO Dario Amodei, and Eleven Labs CEO Mati Staniszewski.