HomeWinBuzzer NewsOpenAI’s GPT-4b Micro AI Model Targets Protein Engineering for Longevity

OpenAI’s GPT-4b Micro AI Model Targets Protein Engineering for Longevity

OpenAI’s coming GPT-4b Micro model applies AI to optimize proteins for better stem cell production

-

OpenAI has been developing an AI model in collaboration with Retro Biosciences, to address inefficiencies in the process of reprogramming adult cells into stem cells.

Early results from laboratory tests indicate that the model, dubbed GPT-4b Micro, could improve the efficiency of Yamanaka factors, proteins critical to stem cell creation, by more than 50 times, reports MIT Technology Review.

Although not officially announced, this project demonstrates OpenAI’s first major step into biological research, offering new possibilities in regenerative medicine.

The collaboration began approximately a year ago when Retro Biosciences approached OpenAI. The startup, based in San Francisco, focuses on extending human lifespan by a decade through advancements in cellular reprogramming.

Related: Google DeepMind Open-Sources AlphaFold 3 to Researchers

“We threw this model into the lab immediately and we got real-world results,” said Joe Betts-Lacroix, CEO of Retro Biosciences, in a discussion reported by MIT Technology Review.

The inefficiency of current methods—where less than 1% of cells are successfully reprogrammed, requiring weeks of effort—was a key motivation for this collaboration. The potential to improve these outcomes could accelerate the development of therapies targeting age-related diseases and tissue regeneration.

The Science Behind Yamanaka Factors and Their Challenges

Yamanaka factors, named after Nobel laureate Shinya Yamanaka, are proteins capable of inducing mature cells to revert to a pluripotent state, allowing them to transform into any cell type.

This reprogramming process underpins advancements in regenerative medicine, from creating replacement tissues to enabling organ regeneration. Despite their promise, the process remains inefficient, slow, and resource-intensive.

OpenAI’s GPT-4b Micro was trained to optimize the performance of these proteins. By analyzing protein sequences and interaction data across species, the model proposes substantial amino acid changes to enhance protein function.

“Just across the board, the proteins seem better than what the scientists were able to produce by themselves,” said John Hallman, an OpenAI researcher. This level of optimization, where up to a third of a protein’s amino acids might be modified, is beyond the reach of traditional methods, which rely on trial-and-error experiments in the lab.

A Unique Application of AI in Biotechnology

GPT-4b Micro represents a distinct approach to AI-driven biotechnology. While Google DeepMind’s AlphaFold focuses on predicting the 3D structures of proteins, aiding researchers in understanding molecular interactions, GPT-4b Micro is designed to improve protein functionality.

This difference highlights the complementary roles these AI tools play in addressing challenges in biological research.

“The model’s ideas were unusually good, leading to improvements over the original Yamanaka factors in a substantial fraction of cases,” Betts-Lacroix explained.

To achieve these results, OpenAI researchers employed a “few-shot” learning technique, where the AI system is guided with a small number of examples before generating optimized designs. This method allows for quick adaptation to specialized problems, such as redesigning proteins for improved performance.

Collaboration with Retro Biosciences and Ethical Considerations

Founded in 2021, Retro Biosciences focuses on advancing cellular therapies and reprogramming techniques as part of its mission to extend human lifespan.

OpenAI CEO Sam Altman, a vocal proponent of AI-driven scientific discovery, personally invested $180 million in Retro, underscoring his belief in the potential of such collaborations.

Altman has previously stated, “Superintelligent tools could massively accelerate scientific discovery and innovation well beyond what we are capable of doing on our own.”

While no financial transactions occurred between OpenAI and Retro Biosciences in this partnership, Altman’s dual role as an investor and CEO has drawn scrutiny. OpenAI clarified that Altman was not directly involved in the development of GPT-4b Micro, emphasizing the project’s focus on advancing scientific knowledge rather than aligning with any specific commercial interest.

This transparency is critical as OpenAI navigates its expanding role in fields beyond general-purpose AI.

Implications for Regenerative Medicine

The advancements enabled by GPT-4b Micro have the potential to address some of the most pressing challenges in regenerative medicine. Enhanced Yamanaka factors could pave the way for more efficient organ regeneration, personalized cell therapies, and the development of artificial tissues.

Harvard University aging researcher Vadim Gladyshev, an advisor to Retro Biosciences, highlighted the broader implications: “[Skin cells] are easy to reprogram, but other cells are not,” he said. “And to do it in a new species—it’s often extremely different, and you don’t get anything.”

OpenAI and Retro Biosciences plan to publish their findings in peer-reviewed journals, providing the scientific community with an opportunity to validate and build upon this research. While GPT-4b Micro is currently a research demonstration and not available for commercial use, its results highlight the transformative potential of AI-driven approaches in biotechnology.

AlphaFold and the Role of AI in Scientific Discovery

OpenAI’s work on GPT-4b Micro complements recent advancements in AI-driven biology, such as DeepMind’s AlphaFold 3, which was recently open-sourced for non-commercial research.

AlphaFold has revolutionized structural biology by accurately predicting the shapes of proteins, enabling breakthroughs in drug discovery and disease research. In contrast, GPT-4b Micro focuses on functional optimization, addressing inefficiencies that hinder the practical application of protein reprogramming.

Markus Kasanmascheff
Markus Kasanmascheff
Markus has been covering the tech industry for more than 15 years. He is holding a Master´s degree in International Economics and is the founder and managing editor of Winbuzzer.com.

Recent News

0 0 votes
Article Rating
Subscribe
Notify of
guest
0 Comments
Newest
Oldest Most Voted
Inline Feedbacks
View all comments
0
We would love to hear your opinion! Please comment below.x
()
x