OpenAI a new goal has appeared: extend human life using AI. An artificial intelligence company has announced that it is working on a new model that attempts to reverse engineer proteins with the help of a scientific startup. OpenAI will follow in the footsteps of Google, which won a Nobel Prize for developing AlphaFold.
Reportedly MIT Technology ReviewOpenAI presented GPT-4b micro, a small language model (SLM) designed for protein engineering. directed. This development, carried out in collaboration with longevity research firm Retro Biosciences, marks the company’s first foray into specialized biological research.
GPT-4b micro aimed at increasing the effectiveness of Yamanaka factorsa set of proteins critical for cellular reprogramming. Factors cause transformation of somatic cells, returning them to a young stem cell state. They have the ability to differentiate into different cell types, which opens up the potential for regenerative medicine, tissue engineering and disease modeling.
How GPT-4b works – a model that can extend human life
According to the researchers, this study differs from traditional methods of predicting protein structure. Instead, Retro uses a language model architecture to suggest modifications to protein sequences.
The purpose of GPT-4b micro is propose modifications of Yamanaka factors to increase the efficiency of their reprogramming.. This is achieved through strategy clue similar to learning from multiple examples, also known as training in several stages. It provides the model with examples of protein sequences and their associated functions, allowing it to extrapolate and generate new sequence variants.
Unlike AlphaFold, which predicts protein folding, GPT-4b micro focuses on sequence manipulation. Model was trained using a dataset covering protein sequences from various species.as well as information about protein-protein interactions. Retro scientists believe that the Yamanaka factors will pave the way for the creation of human organs and replacement cells.
Preliminary data suggests Significant increase in restructuring efficiencywith a reported increase of more than 50 times for some modified factors. Although early results are promising, further testing and peer review are still needed.
The development of GPT-4b micro represents OpenAI’s first attempt to create custom-built models for biological research. Yes, ok the model is not yet publicly availableOpenAI and Retro Biosciences expect to publish their findings in the near future.
Source: Hiper Textual
