Artificial intelligence (AI) technologies continue to revolutionize many fields, such as writing poetry and computer programs, and even creating images of teddy bears and videos of cartoon characters that look like something out of a Hollywood movie. Now, these technologies have the capacity to generate blueprints for microscopic biological mechanisms that can edit the human genome. Artificial intelligence, which has previously discovered medicine and created artificial memories, is now producing molecular mechanisms that can change the human genome
Profluent, a startup based in Berkeley, California, introduced this new technology in a research paper published on April 22, 2024. Profluent’s technology is based on methods guided by ChatGPT, the online chatbot that kicked off the artificial intelligence boom in 2022. The company plans to present this work at the annual meeting of the American Society for Gene and Cell Therapy in May 2024.
Profluent’s Research is Open Source
Just as ChatGPT learns to create language by analyzing Wikipedia articles, books and chat logs, Profluent’s technology analyzes large amounts of biological data, including the microscopic mechanisms used to edit human DNA. From this analysis, new gene editors are created. These gene editors are based on the Nobel Prize-winning CRISPR technology. CRISPR has revolutionized medicine by altering genes that cause inherited diseases such as sickle cell anemia and blindness.
James Fraser, professor of bioengineering and therapeutic sciences at the University of California, San Francisco, who read Profluent’s research paper, notes that the methods developed by artificial intelligence did not exist until today. He emphasizes that AI has learned these methods from nature, but what it produces is new.
Profluent has open-sourced one of these new AI-generated gene editors, OpenCRISPR-1. This allows academic labs and companies to experiment with the technology for free. While AI researchers often open source their AI systems, it is rarer for biology labs and pharmaceutical companies to open source such inventions.
Profluent’s Technology Has Great Potential to Treat Diseases
Ali Madani, CEO of Profluent, explains that these artificial intelligence models work by learning from sequences. Whether these are sequences of characters or words, computer code or amino acids. The company’s technology is driven by an AI model that learns from amino acid and nucleic acid sequences that describe the microscopic biological mechanisms scientists use to regulate genes. By analyzing the behavior of CRISPR gene editors taken from nature, this model learns how to generate entirely new gene editors.
But because these synthetic gene editors have not yet undergone clinical trials, it is unclear whether they can match or surpass CRISPR’s performance. Fyodor Urnov, a gene editing pioneer and scientific director at the Institute for Innovative Genomics at the University of California, Berkeley, says there is no shortage of naturally occurring gene editors, the challenge is the cost of putting them through preclinical studies such as safety, manufacturing and regulatory reviews.
Profluent’s technology could have great potential in the treatment of diseases. Generative artificial intelligence systems tend to evolve rapidly as they learn from ever-increasing amounts of data. If this technology continues to evolve, it could eventually allow scientists to edit genes in much more precise ways, creating a world where medicines and treatments are personalized much faster than today. The potential of AI technology to edit the human genome points to an exciting future for medicine.
The work of innovative startups like Profluent opens the door to a new era in the evolution of gene editing technologies. It should be noted that these technologies need to be carefully considered from a safety and ethical perspective, but the promising possibilities offered by scientific advances cannot be ignored. It should be noted that the research has not yet been peer-reviewed.
It is now published in the preprint in bioRxiv. You can review the research by clicking the;
