Profluent unveils Protein2PAM

New research validates the power of machine learning models to design bespoke biological systems at scale

Our friends at Profluent, the AI-first protein design company, have unveiled Protein2PAM, the first in a series of AI models that will broadly focus on protein-DNA interactions.

Trained on large, bioinformatically curated databases of mined biological sequences, Protein2PAM demonstrates the power of Profluent’s language models to reprogram complex biological systems without the need for wet laboratory evolution or structural modeling.

Early results demonstrate a first proof-of-concept for gene editing, one of the many potential applications of the team’s broader platform.

Protein2PAM confirms that it is possible to use LLMs to scale the precision engineering of proteins without expensive wet lab processes and equipment.

As an initial use case, Profluent has focused on making CRISPR gene editing more versatile. Currently, CRISPR can only edit DNA at specific locations that contain a specific DNA sequence known as a protospacer adjacent motif (PAM). While these markers help ensure accuracy, they also limit where CRISPR can be used. This is a major roadblock for developing treatments for many genetic conditions that could otherwise potentially be helped by gene editing. Previous methods to work around this limitation were either too slow or risked making mistakes.

Profluent trained Protein2PAM on a dataset of protein-PAM associations consisting of 45,816 pairs of Cas proteins and PAMs, representing an approximately 200-fold increase compared to the largest publicly available dataset.

Without any laboratory evolution or structural modeling, Profluent’s AI platform engineered Cas variants that recognize new PAMs and the ability to cut DNA 50 times faster than their natural counterparts. These results were validated through wet lab assays performed in collaboration with the Kleinstiver Lab at the Center for Genomic Medicine at Massachusetts General Hospital (MGH) and the Department of Pathology at MGH and Harvard Medical School.

To make this work easily and freely accessible worldwide, Profluent has created the Protein2PAM Server. This tool enables users to explore and optimize PAM selectivity for their Cas enzyme of choice.

If you’re new to Profluent - this is just the latest in a series of impactful research breakthroughs, including OpenCRISPR - the world’s first open source AI-generated gene editor. You can read more about it on Air Street Press and in the New York Times.

You’ll be hearing more from Profluent soon. I’m planning to interview Ali Madani, the company’s co-founder and CEO, in the near future for Air Street Press. Meanwhile, the team will be at JP Morgan's 43rd Annual Health Care Conference in San Francisco next week.