The Krauthammer Lab at DQBM contributed to a new study in Nature Communications that systematically uncovers Cas9 PAM diversity through large-scale metagenomic mining and machine learning . The team built CRISPR-PAMdb, linking 8,003 Cas9 clusters to inferred PAM profiles derived from 3.8 million microbial genomes and 7.4 million mobile genetic elements.
To extend coverage beyond alignment-based inference, the researchers developed CICERO, a protein language model-based predictor that generated PAM profiles for more than 50,000 additional Cas9 variants. High-confidence predictions reached accuracies up to 0.95 on experimentally validated datasets.
Together, these resources provide a scalable framework for discovering and prioritizing Cas9 orthologs with diverse PAM compatibilities, enabling more flexible and precise genome editing applications.
Publication: https://doi.org/10.1038/s41467-026-69098-5