Finding optimal targets for genome editing or regulatory interference using CRISPR/Cas or zinc-finger nuclease systems.
Two questions GT-Scan can help answer are:
- What are the best candidate targets in my genomic region of interest?
- How many potential off-targets does the target I am using have in the genome?
Candidate targets are places in your genomic region that match a user-defined target rule. Potential off-targets are other places in the genome that are the same as the candidate target except for up to three mismatches, and that match a user-defined off-target filter.
GT-Scan finds each candidate target in your region of interest and reports:
- the number of potential off-targets it has in the genome,
- the genomic sequence and location of each potential off-target, and
- how similar each potential off-target is to it.
GT-Scan sorts candidate targets by how many potential off-targets they have, rather than using a (possibly obscure) ranking formula. It sorts successively by the number of potential off-targets at four levels of increasing severity: 0, 1, 2 or 3 mismatches.
For each candidate target, GT-scan describes each of its potential off-targets reporting:
- the total number of mismatches,
- the number of low-specificity mismatches,
- the number of high-specificity mismatches,
- the positions of the mismatches, and
- a link to the potential off-target in the Ensembl genome browser.
This information can help you decide on the best candidate target (Question 1) or evaluate a given target (Question 2).
GT-Scan lets you define a target rule that specifies
- the length of candidate targets,
- required letters in candidate targets (e.g., the 3'-NGG for CRISPR/Cas systems), and
- the level of binding specificity (zero, low or high) at each position.
GT-Scan also lets you define an off-target filter. Potential off-targets that don't match the filter are ignored. For example, you can require off-targets to match either 3'-NAG or 3'-NGG for CRISPR/Cas systems.
GT-Scan can also ignore potential off-targets with more than a maximum number of high-specificity mismatches that you specify.