Tech | Visa | Scholarship/School | Info Place

EvolutionaryScale, backed by Amazon and Nvidia, raises $142 million for protein-generating AI

A relatively new startup called EvolutionaryScale has received a slew of funding to build artificial intelligence models that generate novel proteins for scientific research.

EvolutionaryScale announced today that it has raised $142 million in a seed round led by former GitHub CEO Nat Friedman, Daniel Gross, and Lux ​​Capital, with participation from Amazon and Nvidia’s corporate venture capital arm NVentures. The company also released ESM3, an AI model described as a “cutting-edge model” for biology that can create proteins for use cases such as drug discovery and materials science.

“ESM3 is a step toward a future in biology where artificial intelligence will be a tool for engineering from first principles, just as we design structures, machines and microchips and write computer programs,” Alexander Rives, co-founder and chief scientist at EvolutionaryScale, said in a statement.

Reeves, along with Tom Secru and Sal Candido, began developing generative AI models to explore proteins at Meta’s AI research lab, FAIR, in 2019. After their group dissolved, Reeves, Secru and Candido left Meta to continue the work they had started.

Characterizing proteins can reveal the mechanisms of disease, including ways to slow or reverse disease, while create Proteins could enable entirely new classes of drugs, tools and therapies. But the current process of designing proteins in the lab is costly, both in terms of computational resources and human resources.

Designing a protein requires designing a structure that can It seems reasonable Performing tasks in vivo or in a product, it then finds protein sequences that are likely to “fold” into a structure (i.e. the sequence of amino acids that make up a protein). Proteins must fold correctly into a three-dimensional shape to perform their intended function.

Rives said ESM3 was trained on a dataset of 2.78 billion proteins and can “reason” about the sequence, structure and function of proteins, allowing the model to generate new proteins in the same way that Google DeepMind’s AlphaFold does. EvolutionaryScale is making the full 98 billion parameter model available for non-commercial use through its cloud-based Forge developer platform and has released a smaller version of the model for offline use.

EvolutionaryScale claims it used ESM3 to generate new variants of green fluorescent protein (GFP), the protein responsible for jellyfish bioluminescence and coral bioluminescence. Its work is detailed in a preprint paper on the company’s website.

Fluorescent protein “esmGFP” created using ESM3 from EvolutionaryScale.
Image Source: Evolutionary Scale

“We’ve been working on this for a long time, and we’re excited to share it with the scientific community and see what they do with it,” Reeves continued.

Of course, EvolutionaryScale isn’t a charity — the company, which has about 20 employees, told TechCrunch that it plans to make money through collaborations, royalties, and revenue sharing. For example, EvolutionaryScale might work with pharmaceutical companies to integrate ESM3 into their workflows, or share revenue with researchers for breakthrough discoveries that are commercialized using ESM3.

To this end, EvolutionaryScale said it will soon make ESM3 and its derivatives available to select AWS customers through AWS’s SageMaker AI development platform, Bedrock AI platform, and HealthOmics services. ESM3 will also be available to select customers using NVIDIA’s NIM microservices and supported by Nvidia enterprise software licenses.

EvolutionaryScale says both AWS and Nvidia customers can fine-tune ESM3 with their own data if they wish.

It may be a while before EvolutionaryScale is profitable. In its promotional materials, a copy of which Forbes managed to obtain last August, EvolutionaryScale repeatedly stresses that it could be a decade before generative AI models can help design therapies. The company will also have to fend off competitors like DeepMind spinoff Isomorphic Labs, which has already signed contracts with big pharma as well as Insitro, publicly traded Recursion, and Inceptive.

EvolutionaryScale’s big bet is to expand its model training to incorporate data beyond proteins and create a general AI model for biotech applications.

A spokesperson for EvolutionaryScale said: “The incredible pace of new advances in AI is driven by ever-larger models, ever-larger datasets, and ever-increasing computational power. The same is true in biology. Over the past five years of research, the ESM team has explored scaling in biology. We found that as language models scale, they lead to a deeper understanding of fundamental biological principles and discoveries about biological structure and function.”

To a journalist, that might sound overly ambitious—but having ambitious investors certainly helps.

#EvolutionaryScale #backed #Amazon #Nvidia #raises #million #proteingenerating

Leave a Reply

Your email address will not be published. Required fields are marked *