HDR Donor Designer — ssODN & dsDNA Repair Templates with Silent PAM Blocking

For a 3′-PAM (Cas9-family) nuclease the blunt cut falls 3 bp inside the PAM-proximal end of the protospacer, so a plus-strand guide cuts at protospacer end − 3 and a minus-strand guide at protospacer start + 2, and the insertion is placed at that cut. Cas12a is refused rather than approximated: its staggered 5′-PAM cut has a different geometry, so supply an explicit edit window for it instead.

The arm length is your call and the tool builds whatever you ask for; the guidance in the field is the usual convention — roughly 30–60 bp per side for a single-stranded oligo donor, 500–1000 bp per side for a double-stranded donor or targeting plasmid (500 is the default). What the tool does enforce is honesty about it: if your target sequence does not extend a full arm past the edit, the arm is truncated and both a warning and a failed gate check say by how much.

Once the donor is repaired in, the allele still matches the guide unless something in the protospacer or PAM has changed — so it can be cut again. When you give a Cas9-family guide whose PAM lies outside the edit window, a single point mutation is folded into the arm that carries the PAM, breaking one of the PAM's fixed bases. It is silent only if a synonymous option exists AND you supplied a CDS reading-frame start; without a frame nothing is checked against a codon, and when no synonymous change breaks the PAM the tool applies a non-silent one and says so. Every candidate mutation it ranked is listed, so you can see what it passed over.

Because a mutation only protects the allele if it survives into the donor, and the donor is only [edit − armLength, edit + armLength]. A PAM further from the cut than the arms reach would be sliced off, so the tool walks its ranked candidate list for one that fits and reports a failure — naming the donor's span — only when none does. Widening the arm length, or picking a guide that cuts closer to the edit, is the fix. It also declines when the edit window already overlaps the PAM, since the edit itself should stop the re-cutting.

The original, unedited target, with the product forced to span the edit site — so the same pair amplifies the wild-type and the knock-in allele and the difference is a size shift or a sequencing read. They come from the same engine as the primer designer, at its default constraints: 18–25 nt, Tm 57–63 °C aiming at 60, GC 40–60%, product 100–1000 bp, and at most five pairs returned. Product lengths are measured before the edit, so add your insert to get the edited band.

No, and it does not pretend to. The gate says so outright: arm-length and design conventions are followed, but HDR rates are cell-type-, locus- and delivery-dependent, and nothing here models them. It also does not check the biology of the edit itself — frame, protein consequence, whether your insert lands on a codon boundary — nor the guide's off-target activity, which is a separate screen.

Yes. Use the explicit-window mode: the window you name is the region being replaced, an empty replacement makes it a pure deletion, and setting end = start − 1 makes it a pure insertion with nothing removed. A guide's coordinates can still be supplied in that mode purely so the PAM gets blocked.

Those two pick and vet the cut; this one designs what repairs it. Find and score candidate protospacers in the CRISPR gRNA Designer, screen the one you like in the Off-Target Checker, then bring its coordinates here to build the donor and the genotyping primers. For base or prime editing — edits that need no donor at all — use the Base Editing Designer or Prime Editing Studio instead.