TRAP-seq

TRAP-seq quantifies CRISPR editing outcomes by capturing and sequencing surrogate target sites paired with gRNA expression cassettes to enable massively parallel measurement of SpCas9, adenine base editor (ABE), and cytosine base editor (CBE) efficiencies.

Key Features:

• Paired gRNA–surrogate integration: Integrates a CRISPR gRNA expression cassette with a corresponding surrogate site to link guide identity to editing outcomes.
• TargetedReporterAnchoredPositionalSequencing: Uses Targeted Reporter Anchored Positional Sequencing to capture and sequence surrogate sites that reflect genomic editing events.
• Quantification of editing outcomes: Measures both indel formation and base editing efficiency for SpCas9, ABE, and CBE.
• Massively parallel / high-throughput scale: Supports large-scale profiling, demonstrated by evaluation of SpCas9 across 12,000 gRNAs in human embryonic kidney cells.
• Nucleotide feature discovery: Identifies sequence determinants of base editing efficiency, including flanking bases, nucleotide motifs, and STOP codon recoding preferences.
• Predictive modeling: Enables development of machine learning models to predict high-efficiency gRNAs for SpCas9, ABE, and CBE with >70% accuracy.
• Data integration: TRAP-seq datasets have been incorporated into The Human CRISPR Atlas for centralized access to editing outcome data.

Scientific Applications:

• High-throughput efficiency profiling: Empirically evaluates editing efficiency of SpCas9, ABE, and CBE across thousands of target sites in cellular contexts.
• Base editor characterization: Measures adenine and cytosine base editing efficiencies and positional preferences across target sequences.
• Guide RNA design optimization: Informs gRNA selection by revealing nucleotide features that correlate with editing performance and by supplying training data for predictive models.
• STOP codon recoding analysis: Assesses patterns and efficiencies of STOP codon recoding mediated by base editors.

Methodology:

Machine learning models were trained to predict high-efficiency gRNAs for SpCas9, ABE, and CBE (achieving >70% accuracy), and TRAP-seq results were integrated into The Human CRISPR Atlas.