Nature Breakthrough: First AI-Designed 'Gene Scissors' Matches Natural CRISPR/Cas, Ushering in a New Era of Genome Editing

However, CRISPR systems have historically been challenging to design because their molecular space is vast and requires optimization across multiple dimensions.

And the emergence of protein language models has givenCRISPR System Brings Customized Opportunities.

Overview of Language Modeling Methods for Designing CRISPR-Cas Systems

Then, by fine-tuning the ProGen2 protein language model, researchers generated 4 million CRISPR-Cas protein sequences. Through rigorous screening and sequence clustering, they found that the diversity of the generated sequences was significantly expanded compared to natural proteins.

For example, for families with very few natural proteins, such as Cas13 and Cas12a, the diversity of the generated sequences increased by 8.4-fold and 6.2-fold, respectively.

Although many CRISPR-Cas proteins have been used for gene editing,Cas9 remains the most widely used editing protein.

Thus, the researchers further utilized a Cas9-specific model to generate 1 million Cas9 protein sequences and constructed a maximum likelihood phylogenetic tree, finding that the generated proteins occupied most of the phylogenetic diversity.

Subsequently, the researchers selected 209 Cas9 analogs for functional validation in human cells and found that some of them...The editing efficiency of the protein is comparable to or even higher than that of SpCas9.

Comparison of Editing Efficiency between OpenCRISPR-1 and spCas9

Researchers conducted detailed editing efficiency and specificity tests on 48 generated Cas9 analogs, finding that many of the generated nucleases exhibited high editing efficiency and specificity, with some even outperforming SpCas9.

Researchers found that, compared with SpCas9, the generated Cas9 analogsImmunogenicityUpper PerformanceLower reactivity.

Experimental validation has shown that the model is capable of generatingHighly functional CRISPR-Cas proteins offer new directions for the development of gene-editing technology.

In the pharmaceutical field, the model generatesCRISPR–Cas proteins can be used to develop safer and more efficient gene therapy solutions.

In the field of agriculture,Gene editing technology can be used to improve crop varieties, enhancing their disease resistance, stress tolerance, and yield. The emergence of new editing tools will provide more options for agricultural biotechnology.

This不得不提到，Who is Profluent?

Profluent is a biotechnology company that uses AI to design proteins. Established in 2023, it secured $9 million in seed funding in the same year. In 2024, the company received an additional $35 million in financial support.

Since its establishment, the company has released multiple AI life science models for generating novel proteins, includingproseLM、Protein2PAM AndProGen3。  

In fact, from the company's perspective, the breakthrough of AI in proteins lies not only in design but also includes a series of related technological advancements, includingManufacturing, Cell Therapy Components and Delivery.

And Profluent's goal is to provide"One-stop"Solution, instead of having customers seek services from 10 different companies.

In April 2024, the company releasedOpenCRISPR-1, currently academic and industry researchers have accessed open-source sequences across various vertical fields, from developing drought-resistant crops to drug discovery.

Profluent Plans to Open Source CRISPR-Cas Atlas to Further Democratize the Gene Editing Field.