CRISPR efficiency is mostly determined before you even open a tube — at the design stage. Good guide RNA design is the difference between a clean knockout in a week and a frustrating screen of colonies that don’t have the edit you wanted.
Step 1: Define your goal
- Knockout: Target an early exon (typically exon 1 or 2) common to all isoforms, ideally upstream of any functional domain.
- Knock-in / precise edit: Position the cut within ~10 bp of the desired edit, since HDR efficiency drops with distance.
- Tagging at termini: Cut as close to the start or stop codon as possible.
- CRISPRi/CRISPRa: Target the promoter region (typically -50 to +300 bp around the TSS for CRISPRi).
Step 2: Pick the right Cas variant
| Cas variant | PAM | Notes |
|---|---|---|
| SpCas9 | NGG | Most common; abundant PAMs |
| SaCas9 | NNGRRT | Smaller protein; AAV-friendly |
| Cas12a (Cpf1) | TTTV | Staggered cut; T-rich genomes |
| SpCas9-NG / xCas9 | NG | Expanded targeting |
Step 3: Use a design tool
- CHOPCHOP — quick, web-based, supports many organisms
- Benchling — integrated with cloning workflows
- CRISPOR — strong off-target scoring; multiple algorithms
- Synthego Knockout Guide Design — optimized for KO experiments
- IDT CRISPR Design Tool — for ordering chemically synthesized sgRNAs
Step 4: Score for on-target efficiency
Tools use machine-learning models (Doench Rule Set 2, Azimuth) trained on large empirical datasets. Higher scores correlate with cutting efficiency, though correlation is imperfect — always validate experimentally.
Step 5: Score for off-targets
Off-target sites have similar but not identical sequences to your guide. Pay particular attention to:
- Off-targets with mismatches only in the 5′ end (more tolerated)
- Off-targets in coding regions or disease-associated genes
- Off-targets in your downstream phenotype (false-positive risk)
Step 6: Avoid common pitfalls
- Polymorphisms. If your cell line has SNPs in the target sequence, the sgRNA may not cut. Sanger-sequence the target region before designing.
- Repetitive elements. Avoid targeting regions with poor mapping.
- Common target across paralogs. If you don’t want to hit a paralog, ensure the guide is unique.
Step 7: Choose a delivery format
- Plasmid encoding sgRNA + Cas9 — cheapest, slower onset, integration risk
- Synthetic sgRNA + Cas9 mRNA — fast, transient, no integration risk
- RNP (sgRNA + Cas9 protein complex) — fastest action, cleanest off-target profile
For most cell line knockouts, RNP delivery via electroporation gives the highest editing efficiency with the lowest off-target burden.
Step 8: Validate
- Surveyor / T7E1 assay: Detects mismatches at the cut site — fast but semi-quantitative
- ICE or TIDE: Sanger-based decomposition of indels in the bulk population
- Amplicon NGS: Most accurate; quantifies all indel types
- Western blot: Confirms protein loss for knockouts
Good sgRNA design takes 30 minutes and saves you weeks. Pick a goal-appropriate target region, use a scoring tool, validate the top 2–3 guides in parallel, and always confirm at the protein level when you can.
