Protein2PAM

Protein2PAM is a deep learning framework to predict the PAM specificity of CRISPR systems from Cas proteins.

Details about the models and their applications can be found in:
Nayfach, S., Bhatnagar, A., Novichkov, A., et al. (2025). Engineering of CRISPR-Cas PAM recognition using deep learning of vast evolutionary data. bioRxiv. https://doi.org/10.1101/2025.01.06.631536.

Protein2PAM is accessible via our webserver at: https://protein2pam.profluent.bio. It was developed at Profluent and code will be made available upon publication.

Models

Below are the full listing of available models utilized in the manuscript and/or webserver:

Model Name	Input Protein/Domain	CRISPR Type	Samples	PAMs from Literature	Used in Webserver?
cas8	Cas8 or Cas10d	Type I	28,410		✅
cas9	Cas9 PI-domain	Type II	15,843	1-9	✅
cas12	Cas12 protein	Type V	1,720	10-21	✅
cas9_full	Cas9 protein	Type II	15,843	1-9	❌
cas9_full_nolit	Cas9 protein	Type II	15,731		❌
cas9_pid_nolit	Cas9 PI-domain	Type II	15,731		❌
cas9_pid_nme	Cas9 PI-domain	Type II	15,843	1-9	❌
cas12_no_lit	Cas12 protein	Type V	1,675		❌

• Model Name should be specified if running Protein2PAM from python
• Samples indicates the number of protein:PAM pairs used for model training
• PAMs from Literature indicates that training samples were supplemented with in vitro determined PAMs from publications; for full listing, see REFERENCES.md
• Used in Webserver? indicates the models that are utilized by the Protein2PAM Webserver
• cas9_pid_nme model was trained using a training dataset in which Protein:PAM samples from Nme orthologs were upweighted

Cite this work

If you use Protein2PAM in your research, please cite the following preprint:

@article{nayfach2025protein2pam,
  title={Engineering of CRISPR-Cas PAM recognition using deep learning of vast evolutionary data},
  author={Stephen Nayfach and Aadyot Bhatnagar and Andrey Novichkov and Gabriella O. Estevam and Nahye Kim and Emily Hill and Jeffrey A. Ruffolo and Rachel Silverstein and Joseph Gallagher and Benjamin Kleinstiver and Alexander J. Meeske and Peter Cameron and Ali Madani},
  journal={bioRxiv},
  year={2025}
  publisher={Cold Spring Harbor Laboratory}
}