What Is Homology Directed Repair in CRISPR: Templates and Outcomes

Homology-directed repair in CRISPR is a template-dependent DNA-repair pathway that copies sequence from a donor into a Cas-induced break. Unlike NHEJ, HDR can install a defined substitution, tag, or cassette when a suitable donor and cell-cycle state are present.

Donor type, homology-arm design, and cell-cycle limits set what an experiment can attempt. HDR does not guarantee a high knock-in rate in every cell type.

How HDR Differs from NHEJ After a CRISPR Cut

Cas nucleases create a double-strand break at a guide-specified site. The cell, not the ribonucleoprotein, decides how that break is repaired. Two pathways dominate the planning conversation for knock-out versus knock-in work.

NHEJ ligates the broken ends, often with a small indel that can shift a reading frame. HDR uses flanking homology to copy the donor sequence across the junction, which is how a SNP, a restriction site, a peptide tag, or a larger cassette can be written in. Alternative end-joining routes exist, but they do not replace the HDR-versus-NHEJ choice when a defined template edit is the goal.

Because NHEJ does not wait for a donor or for S phase, it often wins the competition at the same break. That competition is a biological limit, not a software setting. Planning an HDR experiment means designing a donor that can be used if the cell enters the right state, then verifying the locus by sequencing, not assuming the intended allele is the majority product.

Donor Templates: ssODN Versus Plasmid Donors

The donor is the template HDR copies. Two practical classes cover most CRISPR knock-in designs: a single-stranded oligodeoxynucleotide (ssODN) and a double-stranded DNA donor, usually a plasmid or a PCR fragment.

An ssODN is a short synthetic strand with homology on both sides of a small edit. It is the usual choice for a base substitution, a few amino acids, a short tag, or a diagnostic restriction site. The oligo is easy to order and does not require cloning a donor plasmid. It is a poor carrier for a long expression cassette, because length and synthesis quality become limiting.

A plasmid or other dsDNA donor carries longer inserts between homology arms. Fluorescent tags, selection markers, and whole cassettes typically need this format. The extra sequence is also extra DNA in the cell, so the design must consider how the donor is delivered and how unintended plasmid insertion will be screened. A PCR dsDNA fragment sits between the two: longer than an oligo, without a bacterial backbone, still requiring clean homology arms.

Neither format guarantees HDR. The oligo can be ignored in favor of NHEJ; the plasmid can integrate at a non-homologous junction. Choose the donor by edit size and screening plan, then confirm alleles rather than counting colonies as knock-ins.

Homology Arms and What They Must Match

Homology arms are the sequences on the donor that match the genomic (or episomal) DNA flanking the break. HDR uses that match to copy the intervening edit. Arms that do not match the actual allele in the cell, including SNPs between the reference and the experimental line, reduce the chance of correct conversion.

Arm length scales with donor class. For ssODNs, each arm is commonly tens of nucleotides, with many small-edit protocols using roughly 30 to 80 bases of homology per side. For plasmid or long dsDNA donors, arms of several hundred base pairs are widely used so the recombination machinery can find a longer stretch of identity. These are design ranges, not a promise that a particular length will convert a given cell line.

Place the intended edit close to the cut. Conversion efficiency tends to fall as the mismatch moves away from the break, so a distant SNP in the donor may never appear in the repaired allele. If the nuclease can recut the repaired sequence, include silent changes that destroy the PAM or the seed while preserving the amino-acid code, so the product is not immediately cut again.

Build the donor from the sequence of the target cell when possible, not only from a generic reference. Then store that donor file with the guide design. Sequence tools help inspect junctions and silent mutations on the plasmid or oligo map before anything is ordered.

Cell-Cycle Limits That Constrain CRISPR HDR

Classic HDR in mammalian cells is tied to replication. In S and G2, a sister chromatid is available and HDR proteins are active in their normal role. In G1, and in many non-dividing or terminally differentiated cells, NHEJ dominates. That is why the same guide and donor can produce useful knock-in clones in a rapidly dividing line and almost none in a resting primary cell.

Cell-cycle synchronization, timed delivery, and pathway-modulating compounds appear in published methods as ways to shift the balance. They remain experimental choices with cell-type-specific behavior. They are not a general override of the S/G2 limit, and they should not be described as a reliable efficiency boost.

Organism and ploidy also bound the experiment. Bacteria, yeast, and some cell lines use homologous recombination more readily than typical human primary cells. Diploid and polyploid genomes require screening both (or all) alleles if a homozygous edit is the scientific goal. HDR at one allele plus NHEJ at the other is a common mixed outcome, not a failed delivery.

Design Checks Before an HDR Experiment

Before ordering a donor, confirm that the intended pathway is HDR and that the reagents can be verified. A knock-out that only needs a frameshift does not need a donor. A knock-in that needs a defined protein tag does.

1. Cut site next to the edit. Choose a guide whose break is close to the substitution or insertion junction.
2. Donor class matched to insert size. Use an ssODN for small rewrites; use dsDNA or a plasmid for long cassettes.
3. Arms identical to the target haplotype. Correct SNPs in the arms so the donor matches the cell.
4. Recut protection. Add silent PAM or seed changes when the edited sequence would still be a substrate.
5. A sequencing readout. Design PCR primers that distinguish HDR, NHEJ, and unedited alleles, including across both junctions for a cassette.

ZettaCRISPR supports guide RNA and sequencing-primer design before the wet-lab step, which is the planning layer around an HDR attempt, not a prediction of knock-in rate. After clones appear, sequence the locus. Junction PCR or a restriction site introduced by the donor can triage wells; it does not replace Sanger or amplicon sequencing of the edited region.

Document the guide, PAM, donor sequence, arm coordinates, and intended silent mutations in the experiment record. A later reviewer should be able to tell whether a clone failed HDR or whether the donor never matched the allele that was actually present.

What Homology-Directed Repair Cannot Promise

HDR is a pathway the cell may use, not a command the nuclease issues. It cannot be counted on in non-dividing cells, it cannot be assumed from a surviving colony, and it cannot be declared complete from a single junction PCR. Off-target cuts, if they occur, are repaired by whatever pathway those sites engage; an HDR donor at the on-target locus does not rewrite those events.

Large insertions remain harder than small substitutions. Toxicity from dsDNA, random donor integration, and mixed alleles all sit outside the simple "add a template" picture. The honest planning statement is: HDR is the pathway that can copy a donor at a CRISPR break, within cell-cycle and template limits, and the allele must be verified.

Keep design files and verification primers with the experiment so a failed clone is interpretable. A cloud-based R&D workspace is one place to hold that context; the biological limits stay the same.

FAQ

What is the difference between HDR and NHEJ in CRISPR?

Both pathways repair the double-strand break that Cas enzymes make. NHEJ joins the ends without a designed donor and often leaves a small indel, which is why it is used to disrupt genes. HDR copies sequence from a homologous donor into the break, which is why it is used for defined substitutions and insertions. NHEJ can act through most of the cell cycle. HDR is strongly favored in S and G2. In the same transfected population, some alleles may be NHEJ products and others HDR products. The experiment plan should therefore include a sequencing assay that can tell those outcomes apart, rather than treating survival or GFP-positive cells as proof of the intended rewrite.

Should I use an ssODN or a plasmid donor for CRISPR HDR?

Use an ssODN when the edit is short: a SNP, a few codons, a small peptide tag, or a diagnostic restriction site. Use a plasmid or long dsDNA donor when the insert is a cassette that an oligo cannot carry, such as a fluorescent protein or a selection marker. ssODNs are simpler to produce and avoid a bacterial backbone; they are not suitable as general vehicles for kilobase-scale inserts. Plasmid donors handle length but require junction sequencing and screening for non-homologous insertion. Match the donor to the edit and to the screening method. Neither format overcomes cell-cycle limits, and neither should be described as guaranteeing knock-in.

How long should homology arms be for CRISPR HDR?

Arm length depends on donor type and edit size. ssODN designs commonly use tens of nucleotides of homology on each side of a small edit, with many protocols in the roughly 30 to 80 base range per arm. Plasmid and long dsDNA donors commonly use homology arms of several hundred base pairs so there is a longer stretch of identity. Longer is not automatically better: arms must match the target haplotype, and the edit should sit near the cut. Treat published ranges as starting design bounds, then verify the allele. Do not treat a chosen length as a predicted conversion rate for a new cell type.

Why is CRISPR HDR limited in non-dividing cells?

HDR that uses a homologous template is coupled to the cell cycle. In S and G2, cells already perform homology-based repair, and the pathway proteins are available. Many non-dividing cells remain in G1-like states where NHEJ is the default response to a double-strand break. Delivering a donor oligo or plasmid does not by itself open that pathway. This is why HDR knock-in is routinely attempted in dividing lines and is difficult in resting primary cells or neurons without specialized methods. If the biology of the project requires a non-dividing system, plan for a different editing strategy or for extensive screening, and do not assume a dividing-cell protocol will transfer.

Can homology-directed repair insert a whole gene with CRISPR?

HDR can copy a large cassette from a dsDNA or plasmid donor if homology arms flank the insert and the cell can use the pathway. A whole gene or a long reporter is therefore a plasmid-donor problem, not an ssODN problem. Size, delivery, and screening become the limits: long dsDNA can be toxic, random integration can mimic a targeted clone, and both junctions must be verified. Some projects use other integration methods when the insert is large. HDR remains a valid mechanism for targeted insertion, but it is not a promise that a particular gene-sized donor will land cleanly at the cut in a given cell line. Sequence both junctions and the cargo before calling the clone correct.

Does an HDR donor prevent CRISPR off-target edits?

No. An HDR donor templates repair at sequences that share homology with the arms, typically the intended locus. Off-target cuts, if they occur, are separate breaks. Those sites are usually repaired by NHEJ or another local pathway, not by your knock-in donor, unless they happen to share enough homology for recombination, which is not the normal design intent. Reducing off-target risk is a guide-choice and nuclease-choice problem, followed by appropriate checking of the edited cells. Guide and sequencing-primer design belong in that planning step. An HDR donor is not a shield against off-target repair, and no donor format should be described that way.

Conclusion

Homology-directed repair in CRISPR copies a donor across a Cas-induced break, which is how defined knock-ins are written. NHEJ competes at the same cut; ssODN and plasmid donors serve different edit sizes; homology arms must match the target; and S/G2 biology limits the pathway. Verify alleles. Do not treat HDR as a fixed efficiency setting. To plan guides and sequencing primers before an HDR experiment, review ZettaCRISPR.