Why it matters: Generative artificial intelligence technology can be more than just a threat to the livelihoods of countless artists and writers. According to Salesforce, a perfectly trained machine learning algorithm can help in the creation of artificial proteins useful for health or environmental protection.
Salesforce Research is working on an innovative bio-application for artificial intelligence algorithms. ProGen is an artificial intelligence model designed to create synthetic proteins. He has been trained in hundreds of millions of protein sequences in text form, resulting in artificial proteins as effective as natural ones in waste disposal.
ProGen was featured in this month’s issue of Nature Biotech, where the researchers also described how the first known three-dimensional structure of an artificial protein was fully developed by an artificial intelligence system. ProGen is a language model that can “generate protein sequences with predictable function in large protein families,” just as ChatGPT can combine different text fragments to get “grammatically and semantically correct sentences in natural language on various topics.”
According to the researchers, the new artificial intelligence model was trained on 280 million protein sequences from more than 19,000 families and “supplemented” with control tags containing the properties of proteins. With a sufficient number of homologous samples from different protein families, ProGen can be further customized to “improve controlled protein generation performance.”
In other words, the generative AI ProGen gives researchers the opportunity to create highly adapted proteins “with desired properties” using a controlled tool. ProGen “learned” the rules of protein synthesis by looking through a database of protein sequences, and was able to generate a set of proteins that the researchers later tested for their actual antibacterial properties in the laboratory.
According to Salesforce, the test results showed that 73 percent of the artificial proteins created by ProGen were “functional” compared to 59 percent of natural proteins.
ProGen shows how scientists can develop a “proactive approach” to protein design, Salesforce reports. The company hopes that in the future this new approach will help accelerate the development of “disease drugs” and enzymes for industrial or environmental applications. Plastic-eating squirrels are another potential use that could change the rules of the game.
Salesforce said it is already using the generative ProGen model to identify “potential treatments” for neurological and autoimmune diseases such as rheumatoid arthritis and multiple sclerosis.