Machine Learning Enhances Precision of Genome Editing Enzymes

Researchers from Mass General Brigham have published a study in Nature demonstrating how scalable protein engineering combined with machine learning can improve genome editing. The team developed a machine learning algorithm called PAMmla, which predicts the properties of approximately 64 million genome editing enzymes. This approach aims to reduce off-target effects, increase editing safety and efficiency, and enable the prediction of customized enzymes for new therapeutic applications.

CRISPR-Cas9 enzymes are widely used for editing genes across genomes; however, they can sometimes act at unintended DNA sites, leading to off-target effects. The study addresses these limitations by using machine learning to enhance the specificity of enzyme targeting. According to corresponding author Ben Kleinstiver, “Our study is a first step in dramatically expanding our repertoire of effective and safe CRISPR-Cas9 enzymes. In our manuscript we demonstrate the utility of these PAMmla-predicted enzymes to precisely edit disease-causing sequences in primary human cells and in mice.” 

A key aspect of CRISPR-Cas9 technology is the enzyme’s ability to recognize and bind to a short DNA sequence called a protospacer adjacent motif (PAM). The researchers used PAMmla to predict the PAMs for millions of Cas9 enzymes and identified novel engineered enzymes with improved on-target activity and specificity. Proof-of-concept experiments in human cells and a mouse model of retinitis pigmentosa showed that these custom enzymes had greater specificity.

Lead author Rachel A. Silverstein stated, “A major outcome of this work is the creation of this PAMmla model that can now be used by researchers to predict customized enzymes that are uniquely tuned for their specific use cases.” The team has also made a web tool available for the scientific community to access the PAMmla model at https://pammla.streamlit.app/.