Plasmid Stability Checker — Find the Repeats Your Construct Can Delete
Locate the exact direct repeats that let a construct recombine away the DNA between them, and build the shortened molecule you would actually recover.
🌐 Nothing you paste is logged or stored — every tool is also callable via REST & MCP, and in bulk from the batch tools
A multi-gene construct that shrinks in culture, or comes back from sequencing missing a cassette, has usually recombined with itself. Two identical stretches pointing the same way are all it takes: the DNA pairs one copy against the other and everything between them leaves as a circle, so you keep a construct with one copy where the map says two. Reusing the same terminator after two coding sequences, or the same promoter twice in a MoClo stack, is the ordinary way to build this in — and it is silent, because the clone grows perfectly well and only sequencing shows it. This finds every exact direct repeat above a length you choose, gives both copies' coordinates, and constructs the molecule the recombination would leave, as a sequence you can paste into a map or hand to the Diagnostic Digest Planner to design a screen against. On a plasmid it also gives the excised circle, because the event makes two molecules and which one you recover depends on which inherited the origin and the marker. What it deliberately will not give you is a deletion rate: that depends on the strain's recombination genotype and growth history as much as on the sequence, so findings are ordered by repeat length and spacer — the two factors that do govern it — with no number attached.
Working on a whole construct rather than one pasted sequence? Open SeqStudio — the full editor, with live feature annotation, plasmid maps, restriction and primer panels, undo/redo, multi-document tabs and GenBank / SnapGene import and export.
How to use the Plasmid Stability Checker tool
- 1Paste the construct into the sequence box and tick Circular if it is a plasmid.
- 2Set the shortest repeat worth reporting — 20 bp is the default, below which recombination has little to work with and chance matches multiply.
- 3Read the pairs, longest and closest first: each gives both copies' coordinates and how much DNA sits at risk between them.
- 4Copy a deletion product to check it against your map, or screen for it with the Diagnostic Digest Planner — and fix the design by swapping one of the two repeats for an equivalent part.
Frequently asked questions
Why does reusing the same terminator cause this?
Because two copies of it are two identical sequences in the same orientation, which is exactly the substrate homologous recombination pairs up. Once paired, the DNA between them is excised as a circle and you keep one copy. The parts themselves are fine — it is having two identical copies in one molecule that is the problem, and it is why part collections ship several interchangeable terminators rather than one good one.
How likely is the deletion actually to happen?
That is not answerable from sequence, and no number is offered here. The frequency depends on the repeat length and spacing, but also on the host's recombination genotype (recA, sbcC and friends), the plasmid's copy number and how long the culture has been growing — none of which is in the file you pasted. What is offered instead is the ordering: longer repeats and shorter spacers recombine more readily, so findings are sorted that way. Treat a long repeat as a design flaw to fix rather than a risk to quantify.
Why are there two products for a plasmid?
Because recombining two direct repeats on a circle resolves one circle into two, not into one shorter one: a circle carrying the arc between the repeats plus a copy, and a circle carrying the rest plus a copy. Both are real molecules. Which one you actually recover is decided by which inherited the replication origin and the selection marker, so both sequences are given and you should check where those features sit before assuming.
Does it find inexact repeats?
No — only exact matches. A pair with a few mismatches still recombines, less readily, and this will not see it. That is a real limit rather than a claim that near-repeats are safe: treat a clean result as “no exact repeat of that length”, not as “no recombination substrate”.
What about inverted repeats?
Different mechanism, different consequence, and out of scope here. An inverted pair is a hairpin template: what it costs you is failed synthesis, polymerase slippage and a sequencing read that dies at the stem — not a missing kilobase. The Construct QC Linter flags local hairpins and the RNA Folding tool gives their energetics.
How does this differ from the Construct QC Linter?
The linter looks for tandem repeats of unit length one to six repeated three or more times — microsatellite-shaped runs that complicate synthesis and sequencing. Neither it nor anything else in the linter reaches a pair of identical forty-base blocks a kilobase apart, which is the plasmid-instability case and the only thing this tool looks at. Run both.
Can I run it from code?
Yes — repeat_instability is on the REST API and the MCP server, it works per-record in batch mode over a plasmid collection, and it carries an objective pass/fail gate (no exact direct repeat at or above your threshold). Its deletion products chain straight into diagnostic_digest to design the screen that would catch one.
More
Related tools
Work backwards from the size you measured: which pairs of priming sites on your template could make a product that big, and can their 3' ends actually extend?
Drop in an unedited control trace and your edited pool's trace and get the indel spectrum, the unedited fraction, and the R² that says whether the decomposition means anything.
Choose one primer pair per target so no two cross-dimerise, every amplicon is separable on the gel, and one annealing temperature serves the whole panel.