CRISPR Guide RNA Design: A Complete Guide for Genome Editing Experiments

Understanding Guide RNA Structure

A CRISPR guide RNA (gRNA) consists of two essential parts: a CRISPR RNA (crRNA) containing the target-specific spacer sequence (17-24 nt), and a trans-activating crRNA (tracrRNA) that binds the Cas protein. In synthetic single-guide RNAs (sgRNAs), these are fused into a single molecule.

The spacer sequence determines target specificity through Watson-Crick base pairing with the genomic DNA. For Cas9, the target must be adjacent to a protospacer adjacent motif (PAM) — NGG for Streptococcus pyogenes Cas9.

gRNA Design Rules by CRISPR System

gRNA Target Selection Strategy

The first step in gRNA design is identifying all possible PAM-adjacent sequences in your target region. For SpCas9, this means scanning for NGG motifs and extracting the 20 bp upstream sequence as candidate gRNAs. Most genes will have dozens to hundreds of candidate gRNAs. The selection process must balance:

• On-target efficiency: Not all gRNAs cut equally well. Efficiency depends on spacer sequence, chromatin accessibility, and local GC content.
• Off-target specificity: Mismatches tolerated by Cas9 can lead to cleavage at unintended sites. The seed region (8-12 bp proximal to PAM) is most sensitive.
• Functional relevance: For gene knockout, target early constitutive exons or essential protein domains to ensure loss of function.

On-Target Efficiency Prediction

Several computational models predict gRNA efficiency:

• Azimuth (Doench 2016): The most widely used scoring model, trained on lentiviral tiling screens. Scores are calibrated from 0-100.
• DeepCRISPR: A deep learning model that learns sequence features automatically without manual feature engineering.
• CRISPRScan: Uses a position-dependent scoring matrix derived from high-throughput screens.

For critical experiments, design 3-5 gRNAs per target and validate each. Even the best prediction models have 70-80% accuracy.

Off-Target Prediction Methods

gRNA Synthesis and Delivery

gRNAs can be delivered as:

• Chemically synthesized crRNA + tracrRNA: Highest quality, most flexible, most expensive. ~$30-50 per gRNA.
• In vitro transcribed sgRNA: Requires cloning a template and T7 transcription. Lower cost for bulk production.
• Plasmid-expressed sgRNA: Cloned into expression vectors (e.g., pX330, lentiCRISPRv2). Best for stable expression in cell lines.

High-Throughput gRNA Library Design

For CRISPR screens, libraries of 10,000-100,000 gRNAs are designed computationally. Key considerations include: uniform representation, minimal off-target effects across all gRNAs, and adequate coverage of the library with 3-6 gRNAs per gene. The GeCKO, Brunello, and Yusa libraries are widely used validated designs.