Restriction site finder
Which enzymes cut, the fragment sizes they give, and a map of the cut positions.
Which enzymes cut, the fragment sizes they give, and a map of the cut positions.
Paste the sequence and the Sites tab scans every enzyme in the table below at once, reporting how many of them cut, where, and what kind of end each cut leaves. TheDigest tab takes one or two enzymes and gives the fragment sizes you would see on a gel. The Map tab draws the cut positions along the sequence. All three read plain sequence or FASTA, and the whole calculation runs in your browser.
Only the first record of a FASTA file is analysed, because a digest is a property of one molecule. Whitespace, line breaks and position numbers are ignored, and U is read as T.
A recognition sequence is written with a caret at the point where the enzyme cuts the top strand, and the second caret marks the bottom strand:
A cut position of 84 means the top strand is cut between base 84 and base 85, so the fragment to its left is bases 1 to 84 and is 84 bp long. Fragment sizes are always counted from the top-strand cut. The Ends column names the overhang the enzyme leaves, which is what decides whether two fragments can be ligated to each other.
On the Digest tab pick one enzyme, or two for a double digest. The cut positions of both are pooled and the fragments between them are listed largest first, the order bands appear in from the top of a gel. A linear molecule cut n times gives n + 1 fragments; a circular one gives n.
Worked example. "Try an example" loads a 904 bp cloning construct: a T7 promoter, a polylinker, the EGFP coding sequence and a T7 terminator. The Sites tab reports that54 of the 153 enzymes cut it, at 130 positions in all, and that 27 enzymes cut exactly once. Switching to Digest with EcoRI and HindIII gives 3 fragments of746, 84 and 74 bp: EcoRI cuts after base 84, HindIII after base 830, and 84 + 746 + 74 adds back up to 904.
Most enzymes read a palindrome and cut inside it. Type IIS enzymes such as BsaI, BsmBI, BbsI and SapI read a sequence that is not a palindrome and cut a fixed distance outside it, which is what makes Golden Gate assembly possible. They are written in the REBASE style:
Because the site is not a palindrome it has a direction, so both strands are searched. A BsaI site on the bottom strand reads GAGACC on the top strand and cuts upstream of where the site is written. The example sequence contains one such site in the T7 promoter region, at the GAGACC in the leader. On a linear sequence a cut that would fall beyond either end is reported as a site but leaves no fragment boundary, because there is no DNA there to cut.
Tick Circular DNA for a plasmid. The sequence is then searched across the join between the last and the first base, so a site that straddles position 1 is found, and the fragment list wraps: a plasmid cut once gives a single linear fragment the length of the whole plasmid, and cut twice gives two. Numbering still starts at base 1 of the sequence as you pasted it, which is the same convention plasmid maps use.
153 enzymes, chosen from the commercial catalogues, with recognition sequences and cut offsets as published by REBASE and New England Biolabs. Isoschizomers that cut the same sequence at different points are listed separately, because the ends they leave differ:
| Enzyme | Site | Ends | Enzyme | Site | Ends |
|---|---|---|---|---|---|
| KasI | G^GCGCC | 4 nt 5' | EcoRI | G^AATTC | 4 nt 5' |
| NarI | GG^CGCC | 2 nt 5' | BamHI | G^GATCC | 4 nt 5' |
| SfoI | GGC^GCC | blunt | HindIII | A^AGCTT | 4 nt 5' |
| PluTI | GGCGC^C | 4 nt 3' | NotI | GC^GGCCGC | 4 nt 5' |
| SacI | GAGCT^C | 4 nt 3' | EcoRV | GAT^ATC | blunt |
| Eco53kI | GAG^CTC | blunt | BsaI | GGTCTC(1/5) | 4 nt 5' |
| AatII | GACGT^C | 4 nt 3' | BbsI | GAAGAC(2/6) | 4 nt 5' |
| ZraI | GAC^GTC | blunt | SapI | GCTCTTC(1/4) | 3 nt 5' |
Ambiguity codes in a recognition sequence are expanded, so AccI (GTMKAC) matches GTAGAC, GTATAC, GTCGAC and GTCTAC, and the IUPAC codes are the same ones theIUPAC expander lists. An ambiguous base in yoursequence never counts as a match.
To design the insert that goes between two of these sites, use thePCR product size tool for the amplicon, thesilent mutation finder to add or remove a site without changing the protein, and the ligation calculator for how much insert to add.
Paste the sequence or choose a FASTA file. The Sites tab scans 153 commercial enzymes at once and reports how many of them cut, where each one cuts, and which ones cut exactly once. The enzymes that do not cut at all are listed under Details, which is the fastest way to pick an enzyme that leaves your insert untouched.
That the top strand is cut between base 84 and base 85, so the first fragment is bases 1 to 84. Fragment sizes are counted from the top-strand cut, which is the convention gels are read with. The bottom strand is cut a few bases away when the enzyme leaves an overhang, and the overhang is given in the Ends column.
Yes. Tick Circular DNA and the sequence is searched across the join as well, so a site that spans the origin is found, and the digest gives one fragment for a single cutter instead of two. A circular sequence cut n times gives n fragments; a linear one gives n plus 1.
Yes, with their real cut offsets. BsaI is written GGTCTC(1/5): it cuts the top strand one base after the recognition sequence and the bottom strand five bases after it, leaving a four-base 5-prime overhang. Because these sites are not palindromic, both strands are searched, and a site on the bottom strand cuts upstream of where it is written.
An ambiguous base in your sequence is never counted as a match, because the enzyme would only cut if the real base happens to fit. Ambiguity codes in the recognition sequence work the other way round and do match, so AccI (GTMKAC) finds GTAGAC, GTATAC, GTCGAC and GTCTAC.
No, apart from DpnI, which is listed because it only cuts GATC when the A is Dam methylated. Plasmid DNA grown in a normal laboratory strain is Dam and Dcm methylated, which blocks ClaI, XbaI, MboI and several others at some sites. Check the enzyme supplier before you trust a digest that looks wrong.