Prediction of sgRNA Off-Target Activity in CRISPR/Cas9 Gene Editing Using Graph Convolution Network

Area of Science:

• Genomics and Bioinformatics
• Molecular Biology
• Computational Biology

Background:

• CRISPR/Cas9 is a key genome-editing tool for gene repair and regulation.
• Off-target cleavage by CRISPR/Cas9 poses a significant challenge in gene editing research.
• Existing deep learning models for off-target prediction are complex and difficult for researchers to implement.

Purpose of the Study:

• To develop a novel, user-friendly graph-based approach for predicting CRISPR/Cas9 off-target efficacy.
• To simplify the process of identifying potential off-target sites for researchers.
• To improve the accuracy and accessibility of off-target prediction in genome editing.

Main Methods:

• A graph-based approach was developed, representing sequences as nodes.
• Link prediction was employed to identify connections between sgRNA and off-target DNA sequences.
• Sequence-derived features were extracted and utilized within the graph model.

Main Results:

• The graph-based method demonstrated high accuracy in predicting off-target gene knockouts.
• The model achieved an area under the receiver operating characteristic curve (auROC) of 0.987 on HEK293 and K562 datasets.
• The approach effectively predicted links between sgRNA and off-target sequences.

Conclusions:

• The novel graph-based method offers a simplified and effective way to predict CRISPR/Cas9 off-target effects.
• This approach is easily understandable and replicable for researchers in the field.
• The findings contribute to safer and more precise genome editing applications.