OpenCRISPR-1 Nuclease

Product Introduction

OpenCRISPR-1 is the world's first gene editor designed entirely by artificial intelligence. The research team at Profluent first systematically mined microbial genomic data totaling 26.2 terabases to construct the CRISPR-Cas Atlas database, containing over 1.24 million CRISPR operons. The Cas9 sequence diversity within this database is more than four times greater than that found in UniProt. Based on this massive dataset, the researchers fine-tuned the ProGen2 protein language model to generate 4.8 million novel CRISPR-associated protein sequences. Through rigorous screening and functional validation, they ultimately identified OpenCRISPR-1 as the top-performing candidate.

OpenCRISPR-1 consists of 1380 amino acids and maintains the typical bilobal architecture framework of type II Cas9 nucleases. However, its amino acid sequence differs from SpCas9 at over 400 positions and shares no more than 86% sequence similarity with any known naturally occurring CRISPR-associated protein. Despite this high degree of sequence divergence, AlphaFold2 structural predictions indicate that it fully retains key functional domains, including the HNH and RuvC nuclease domains. OpenCRISPR-1 recognizes the same 5'-NGG-3' PAM sequence as SpCas9 and can directly use standard SpCas9 sgRNAs, making it a plug-and-play replacement for existing SpCas9-based approaches.