How to Predict Restriction Digest Fragment Sizes

You’ve cloned something, and before you sequence it you run a quick restriction digest to check it’s right. The gel shows a ladder of bands — but are they the right bands? You don’t have to guess: if you know the sequence, you can predict the exact fragment sizes ahead of time and just confirm the pattern. Here’s how the prediction works.

Step 1: find where the enzyme cuts

A restriction enzyme binds a specific short recognition sequence and cuts there. EcoRI recognises GAATTC and cuts between the G and the first A (written G^AATTC); SmaI recognises CCCGGG and cuts dead centre (CCC^GGG). To predict a digest you scan the sequence for every occurrence of the site and record the cut position — the base after which the top strand is broken.

Some enzymes have recognition sites with ambiguity: HinfI cuts GANTC, where N is any base. A good calculator expands those IUPAC codes so GAATC, GACTC, GAGTC and GATTC all count as sites.

Step 2: the gaps between cuts are the fragments

Once you have the cut positions sorted along the sequence, the fragments are simply the distances between consecutive cuts. Take a 54 bp fragment cut at positions 1, 7, 13 and 43:

• start → 1 = 1 bp
• 1 → 7 = 6 bp
• 7 → 13 = 6 bp
• 13 → 43 = 30 bp
• 43 → end (54) = 11 bp

That’s five fragments — 30, 11, 6, 6 and 1 bp — and they add back up to 54. The restriction digest calculator does exactly this for the enzymes you select, and shows each enzyme’s site count so you can spot single-cutters at a glance.

Step 3: linear or circular? It changes the count

This is the part people trip over. The same cut positions give a different number of fragments depending on the DNA’s shape:

• Linear DNA (a PCR product, a gel-purified insert) has two free ends, so N cuts give N+1 fragments. Three cuts → four bands.
• Circular plasmid wraps around with no free ends, so N cuts give exactly N fragments. A single cut doesn’t fragment it at all — it just linearises it into one band at the full plasmid length.

So a plasmid that shows one band after a digest has a single site for that enzyme — a handy way to linearise a vector. Toggle “circular” in the tool when your DNA is a plasmid.

Sticky ends vs blunt ends

Whether an enzyme leaves a sticky or blunt end depends on where within the site it cuts each strand. EcoRI cuts off-centre and leaves a four-base 5′ AATT overhang; any two EcoRI fragments have complementary overhangs and can be ligated together, which is what makes sticky-end cloning directional and efficient. SmaI cuts straight across the middle and leaves a blunt end — ligatable to any other blunt end, but with no overhang to enforce orientation.

What the prediction can’t tell you

A sequence-only prediction assumes every site is cut, completely and cleanly. Real digests have wrinkles:

• Methylation — Dam (GATC) and Dcm (CCWGG) methylation in most lab E. coli strains blocks some enzymes from cutting overlapping sites.
• Star activity — under high glycerol, wrong buffer or too much enzyme, some enzymes cut relaxed or extra sites.
• Partial digestion — too little enzyme or time leaves some sites uncut, adding larger “partial” bands.
• Double digests — two enzymes need a shared buffer where both are active, or a sequential digest.

Treat the predicted fragments as your expected pattern, then confirm on the gel. For the reverse-complement strand or to check an insert’s orientation, the reverse complement tool pairs well with this one — and everything runs in your browser, so an unpublished construct is never uploaded.

---

Recognition sites and cut positions here follow standard NEB/REBASE values for 26 common enzymes. The tool predicts a digest from the DNA sequence alone; it does not model methylation, star activity, partial digestion or buffer compatibility. Confirm important results experimentally.