Written and maintained by CASRAI Editorial Board
Last updated
Short answer
CRISPOR is a free web tool, at crispor.org, that takes a genomic sequence, a genome and a Cas nuclease and returns candidate guide RNAs with separate scores for predicted efficiency, specificity and out-of-frame outcome, plus the oligo and primer sequences needed to clone and test them. Its authors state that the predictive value of the scores is not very high, so treat the ranking as a shortlist, not a verdict.
Definition and scope
CRISPOR was described by Haeussler and colleagues in 2016 (Genome Biology 17:148) and updated by Concordet and Haeussler in 2018 (Nucleic Acids Research 46(W1):W242-W245). The 2018 paper reports support for SpCas9, SaCas9 and Cpf1 (Cas12a) and more than 150 genomes added in the two years before publication. It sits at the design step of a knockout workflow; see sgRNA design for the underlying principles and gene knockout for the wider experiment.
How CRISPOR is used in practice
The following summarises the workflow described in the 2018 paper. It does not replace the tool’s own manual.
- Paste the target sequence, typically an exon, in genomic form rather than cDNA.
- Choose the genome and the nuclease.
- Review the guide list. By default guides are sorted by specificity and coloured green, yellow or red.
- Check each guide’s off-target list (up to four mismatches, with CFD scores) and the warnings for high or low GC content and for TTTT runs. Known SNPs can be masked with N.
- Use the cloning and PCR primer output: overlapping oligos for guide expression and flanking genomic primers for validating the edit, including primer pairs for off-target sites. CRISPResso-formatted output is available for analysing sequencing results.
For many guides at once, CRISPOR Batch takes a gene list and produces oligos formatted for pool suppliers. According to the 2018 paper, validated libraries exist only for human and mouse, and a saturating mutagenesis mode is also offered. The code is open source and a command-line version exists. A precomputed CRISPR track (exons plus or minus 200 bp) is available in the UCSC Genome Browser.
Reading the three scores
CRISPOR reports efficiency, specificity and out-of-frame scores separately. The 2018 paper notes that a guide can be predicted to be very specific but not very efficient, or the reverse, so a single combined rank is not assumed. Specificity is one column to read, not the whole decision. Which efficiency model to read depends on how the guide is expressed:
| Guide expression system | Efficiency score the authors recommend | Source |
|---|---|---|
| U6 promoter (expressed in cells) | Doench 2016 / Fusi | Haeussler et al. 2016; Concordet and Haeussler 2018 |
| T7 in vitro transcription | Moreno-Mateos | Haeussler et al. 2016; Concordet and Haeussler 2018 |
Where it matters for lab decisions: evidence on score limits
- Efficiency prediction has a ceiling. Haeussler et al. (2016) put the correlation upper limit for any on-target predictor at about 0.7, and reported Spearman correlations of 0.341-0.436 on independent data.
- Off-target scoring. Across eight SpCas9 off-target studies, the CFD score reached an AUC of 0.91 versus 0.87 for the MIT score. A minimum CFD of 0.023 removed 57% of false positives while losing 2% of true positives.
- Mismatch search depth. 88.4% of validated off-targets had four or fewer mismatches, which is why that is the default search depth. Insertions and deletions are not allowed because they would change only 2 of 225 sites.
- Most predicted off-targets are not real. The 2018 paper cautions that most predicted off-targets are mostly false positives.
- GC content. The 2016 paper advises avoiding GC above 75%.
- Testing more than one guide pays off. In a zebrafish example, choosing the top two guides per locus cut the number of injections from 104 to 22, and the top two found a working guide at 9 of 11 loci.
The same paper compared predicted sites for an EMX1 guide: 1,288 with CRISPOR (Cas-OFFinder based search) versus 334 on the MIT site, and the MIT site missed 5 of 15 validated off-targets. If you use the MIT specificity score (0-100, highlighting 50 and above), the authors suggest a threshold of roughly 70-80.
Common misunderstandings
- Green means it will cut. The colours reflect the specificity ranking, not a guarantee of editing efficiency.
- A short off-target list means a safe guide. The list is a prediction up to four mismatches, not experimental off-target detection.
- One score fits all workflows. The right efficiency model depends on whether the guide is expressed from U6 or transcribed in vitro with T7.
- The -GG rule decides everything. The 2016 paper describes the Wong score and the -GG rule as precise but low in recall, useful as tie-breakers rather than filters.
Related reading: Cas12a for the Cpf1 option, CRISPR screening for pooled library work, and Primer3 and NCBI BLAST for primers for primer design and specificity checks beyond CRISPOR’s built-in primers.
Frequently asked questions
What is CRISPOR used for?
It ranks candidate guide RNAs for a target sequence and outputs the oligos and validation primers needed to clone and test them, for SpCas9, SaCas9 and Cpf1.
Is CRISPOR open source?
Yes. The 2018 paper states the code is open source and that a command-line version exists, alongside the web tool at crispor.org.
Which nuclease and genomes does it support?
The 2018 paper lists SpCas9, SaCas9 and Cpf1, and more than 150 genomes added in the preceding two years. The current genome list is on the tool itself.
Can I trust the efficiency score?
Use it to compare guides, not to promise results. Published correlations on independent data were 0.341-0.436, and the authors’ zebrafish example supports testing more than one guide.
Sources
- Concordet J-P, Haeussler M. CRISPOR: intuitive guide selection for CRISPR/Cas9 genome editing experiments and screens. Nucleic Acids Res 2018;46(W1):W242-W245. doi:10.1093/nar/gky354. PMC6030908 (read 2026-11-08).
- Haeussler M, et al. Evaluation of off-target and on-target scoring algorithms and integration into the guide RNA selection tool CRISPOR. Genome Biol 2016;17:148. doi:10.1186/s13059-016-1012-2. PMC4934014 (read 2026-11-08).
