🧬 1. CRISPR Basics
CRISPR-Cas9 is a programmable nuclease that cuts DNA at a target site defined by two
components:
- Guide RNA (gRNA / spacer) — 20 nt sequence complementary to the target DNA.
- PAM — Protospacer Adjacent Motif, a short sequence (e.g.,
NGG for SpCas9)
that Cas9 requires immediately 3' of the guide-binding site.
Key rule: No PAM = no cut. A perfect 20-nt match without a PAM will not be cleaved.
🎯 2. gRNA Design — Result Interpretation
Input format
Provide a DNA sequence (A/C/G/T) plus a PAM pattern. The tool scans both strands and
returns all possible 20-nt guides followed by a valid PAM.
Columns explained
| Column | Meaning | Preferred value |
|---|
| Guide | 20-nt spacer sequence | Unique in genome |
| PAM | Protospacer Adjacent Motif | Matches chosen Cas variant |
| Strand | + (sense) or − (antisense) | Both are usable |
| Position | 1-based start in target | — |
| GC% | G+C content of the guide | 40–60% |
| Score | Heuristic on-target quality (0–100) | ≥ 80 |
How the score is calculated
- Start at 100.
- −30 if GC < 30% or > 70%; −15 if outside 40–60%.
- −25 if the guide contains
TTTT (Pol III termination signal).
- −15 if it contains
GGGG (secondary structure risk).
- −20 if it can self-anneal (self-complementarity).
Score interpretation
| Score | Rating | Recommendation |
|---|
| 80–100 | Excellent | Top priority — use if off-targets are clean. |
| 60–79 | Acceptable | Use if no better option; validate experimentally. |
| < 60 | Poor | Avoid — high chance of low editing efficiency. |
Worked example
Suppose a candidate GCCATTGTAATGGGCCGCT + GAA scores 85
with GC = 55%. Interpretation:
- GC within optimal range → good stability.
- No poly-T, no poly-G, no self-complementarity → no penalties.
- This guide is a strong candidate for experimental validation.
Tip: Always design 2–3 independent guides per target gene. Even a perfect
in-silico score can fail in the lab.
🔎 3. Off-Target Analysis — Reading the Results
The seed region
The last 10–12 nt at the 3' end of the guide (adjacent to the PAM) — the
seed region — is the most critical for Cas9 binding. Mismatches here usually
abolish activity; mismatches at the 5' end are usually tolerated.
Result interpretation
| Off-target seed matches | Meaning | Action |
|---|
| 0 hits |
No seed match in the searched reference → very low off-target risk. |
Proceed with this guide. |
| 1–2 hits |
Some potential off-targets — inspect the context. |
Verify if the site is in a coding region; consider alternatives. |
| ≥ 3 hits |
High off-target risk. |
Choose a different guide if possible. |
What the columns mean
- Guide — the query guide used in the search.
- Seed region — the last 12 nt used for matching.
- Position in reference — 1-based coordinate of the match.
- Context (±10 nt) — surrounding sequence so you can judge whether the site
looks like a real protospacer (with its own PAM).
Important limitation: This is a simplified seed-based scan, not a full genome-wide
alignment. For real experiments, always confirm with validated tools:
Cas-OFFinder, CRISPOR, Benchling, or CHOPCHOP.
✅ 4. Target-Site Verification
Purpose
Confirms whether a chosen guide + PAM exists in a given target sequence. Useful before
ordering oligos, or to troubleshoot a failed experiment.
Match types
| Match type | What it means | Verdict |
|---|
| exact |
Full guide + PAM present on the + strand. |
✔ Target site is valid. |
| exact (reverse complement) |
Full guide + PAM present on the − strand. |
✔ Valid (CRISPR works on both strands). |
| guide only (PAM missing) |
The 20-nt guide exists but no PAM follows. |
✗ Cas9 will not cut this site. |
| No match |
Guide not found in target. |
✗ Wrong sequence or wrong guide. |
Worked example
Guide = GCCATTGTAATGGGCCGCT, PAM = GAA, target =
ATGGCCATTGTAATGGGCCGCTGAAAGGGTGCCCGATAG.
- Full sequence
GCCATTGTAATGGGCCGCTGAA is present starting at position 3.
- Match type: exact, strand +.
- Interpretation: the target site is valid — proceed to design experiments.
🚀 5. Recommended Workflow
- Obtain target sequence (gene of interest, exon, promoter, etc.) from NCBI or Ensembl.
- Tab 1 — Design: run gRNA design, filter for GC 40–60% and Score ≥ 80.
- Tab 2 — Off-target: check the top 3–5 candidates against a reference; prefer 0–1 hits.
- Tab 3 — Verify: confirm the chosen guide + PAM exists in the target.
- Cross-validate with CRISPOR / CHOPCHOP to catch what the heuristic misses.
- Order 2–3 guides per target and test in parallel.
- Validate editing by Sanger sequencing, T7E1 assay, or NGS.
⚠️ 6. Common Pitfalls & Best Practices
Perfect score ≠ perfect guide: In-silico scores predict but do not guarantee
editing efficiency. Chromatin state, methylation, and delivery method all matter.
Off-target ≠ only seed: Real off-targets can exist with mismatches outside the
seed. Use validated aligners for critical work.
Poly-T warning: Guides containing TTTT can terminate U6-driven
transcription → poor gRNA expression.
SNPs in the target: Population SNPs inside the guide or PAM can abolish cutting
in some individuals. Check dbSNP for your target region.
Best practice: Design 2–3 independent guides, use at least one with a high score
AND a unique seed region. Validate editing by sequencing.
Design checklist
- ✔ GC% between 40–60%
- ✔ No poly-T (TTTT) or poly-G (GGGG)
- ✔ Score ≥ 80
- ✔ 0–1 off-target seed matches in reference
- ✔ Target site confirmed by verification step
- ✔ Cross-validated with a second tool
- ✔ 2–3 guides ordered per gene