Precise, predictable genome integrations by deep-learning-assisted design of microhomology-based templates

Area of Science:

• Genetics and Genomics
• Molecular Biology
• Bioengineering

Background:

• CRISPR-based genome editing offers powerful tools but precise DNA integration and editing remain challenging due to limited control over cellular repair mechanisms.
• Existing methods often result in unpredictable outcomes, including unwanted deletions and insertions, hindering their application in research and therapeutics.

Purpose of the Study:

• To develop a predictable and controllable method for precise CRISPR-based DNA integration and editing.
• To improve the efficiency and accuracy of gene insertion and modification using sequence-context-specific repair strategies.

Main Methods:

• Utilized deep learning models to predict DNA repair outcomes at the genome-cargo interface based on sequence-context rules.
• Designed and implemented base-pair tandem repeat repair arms that match microhomologies at double-strand breaks.
• Validated the strategy across diverse cell types and organisms, including HEK293T cells, Xenopus, and mouse brains, for both germline and somatic applications.

Main Results:

• Demonstrated predictable DNA repair and precise integration of genetic cassettes at 32 loci in HEK293T cells.
• Achieved germline-transmissible transgene integration and successful endogenous protein tagging in Xenopus and mouse brains.
• Showcased scarless single-nucleotide and double-nucleotide edits using optimized repair arms and oligonucleotide templates in vitro and in vivo.
• Developed Pythia, a design tool to facilitate precise genomic integration and editing.

Conclusions:

• Precise CRISPR-based DNA integration and editing are achievable through predictable, sequence-context-specific repair mechanisms.
• The developed strategy significantly enhances the accuracy and efficiency of gene editing, reducing unwanted byproducts.
• This approach holds broad potential for experimental genomics, gene therapy, and synthetic biology applications across various cell types and organisms.