OpenAI and Retro Biosciences have partnered to advance stem cell production technology.
OpenAI’s new language model, GPT-4b micro, assists scientists in extending human lifespan by enhancing stem cell production, as reported by MIT Technology Review. This collaboration with Retro Biosciences focuses on protein re-engineering to improve stem cell generation.
Stem cells are vital in regenerative medicine because of their ability to transform into various cell types, offering potential treatments for age-related illnesses.
Retro Biosciences, established in 2021, aims to extend human lifespan through cellular reprogramming. OpenAI CEO Sam Altman invested $180 million in the company in 2022.
GPT-4b micro analyzes ways to modify Yamanaka factors—proteins that convert adult cells into stem cells. Initial tests show the model’s redesigned proteins are over 50 times more effective at inducing stem cell production than naturally occurring ones.
Trained on extensive biological data from diverse species, the model predicts protein structures and interactions more precisely than conventional methods.
In a Friday article, MIT Technology Review quoted OpenAI developer John Hallman: “just across the board, the proteins seem better than what the scientists were able to produce by themselves.”
Currently in the research phase, GPT-4b micro is not publicly available. OpenAI plans to publish the results for peer review later.
Another project developer, Aaron Jaech, told MIT Technology Review the project solidifies OpenAI’s role in scientific research. He noted it’s unclear “whether those capabilities will come out to the world as a separate model or whether they’ll be rolled into our mainline reasoning models.”
Google’s AlphaFold, introduced in 2018, is an AI model that determines complex 3D protein structures with high accuracy in minutes, according to Google DeepMind.
While in a related field, GPT-4b micro employs different principles to help researchers effectively re-engineer specific proteins.
Numerous research teams globally are using AI to create new treatments. A recent Nature study suggested AI-designed proteins can neutralize deadly snake venom.
In 2022, researchers from the University of Washington School of Medicine and Harvard University trained image-generating AI models to design new proteins for vaccines, cancer treatments, and other applications.
“The proteins we find in nature are amazing molecules, but designed proteins can do so much more,” stated David Baker, a biochemistry professor at UW Medicine and senior author on the study, who later received the 2024 Nobel Prize in Chemistry for computational protein design.