Selecting edited clones without fooling yourself

A resistant colony is evidence that a selectable gene was expressed long enough to matter. That is a smaller fact than an edited allele. Most of the ways to fool yourself in clone picking are words: positive, clonal, null, homozygous, applied to a well that was never one cell and never sequenced. This glossary is those words, tied to the decisions they are allowed to trigger. The mechanism you think you selected for is in how CRISPR-Cas9 editing works in research. The assays that replace the adjectives are in checking whether a genome edit worked.

Plasticware, selective agents and PCR reagents are catalogue classes in the molecular biology catalogue. Specify the genotype you will demand when you open the quote request.

A vocabulary that does not smuggle in a sequence

A pool is every cell that shared a well after delivery. It is a mixture until proven otherwise. A clone is a culture expanded from one cell. The word is a claim about history. It is earned by a method that plausibly deposited one cell, plus a genotype consistent with one genome, plus a later check that the culture did not become mixed again.

An allele is one copy of the locus. Zygosity words assume you know how many copies the parent had. Heterozygous means the copies you accounted for are not all the same. Homozygous means every copy you accounted for carries the same allele. It does not mean the gene is absent, and it does not survive contact with an aneuploid line whose copy number you never measured.

A reporter or a resistance marker is a protein or a cassette you can see or select. It tracks itself. Limiting dilution is a statistical plating method. Single-cell sorting is a cytometer claim that one event was deposited. Both produce wells that are empty, clonal, or secretly double. Passenger mutations are changes that rode along with the clone, including off-target cuts and spontaneous variants. They are not removed by picking a tidy colony.

How a well becomes a candidate, and how it stays a mixture

Start from a pool you have already shown is edited at some frequency worth chasing. Cloning a pool in which one cell in five hundred carries the allele is a search dressed up as a method. Improve the guide or the delivery before you fill plates. The comparison between staying in the mixture and leaving it is in editing pools versus single-cell clones.

If you dilute by hand, calibrate the count, plate a density that makes doublets uncommon, and accept empty wells as the sign that you were near one cell. If you sort, use a gate that excludes doublets as far as the instrument allows, and still assume some wells received two cells. Culture conditions for that outgrowth are the same subject as mammalian cell culture for research labs. A well that grows in three days when neighbours take three weeks may be the unedited parent outcompeting a sick edited cell that was also in the well.

Mark the well only after a genotype. PCR primers from NCBI Primer-BLAST should sit outside the scar. Sequence the product. A clean trace through the cut, different from the parent, can support a single allele or identical alleles. A trace that splits at the cut supports two sequences. Split traces are expected in a diploid heterozygote and also expected in a mixed well. If you need to know which, separate the alleles by cloning the amplicon or by reading the well deeply, and reclone if the culture history was weak.

Drug selection comes before or after that split, and it answers a different question. The logic of selection markers in microbial cloning is a useful analogy, not a protocol to copy into human cells: antibiotic selection and blue-white screening. In mammalian work the antibiotic often selects for the editor plasmid. Keeping the drug on while you want the nuclease gone selects for cells that retained the plasmid. Remove the editor before you call the clone stable, then confirm the locus again. Cells can be recut after you thought the experiment was over.

When the trace is already untrustworthy

If the unedited parent shows the same scar as the clone, stop. You are looking at contamination or at a parental variant you mistook for an edit. Sequence the parent on the same assembly you designed against, from Ensembl or GenBank.

If half the wells show an identical rare indel, suspect a PCR contaminant from one successful clone, not a miracle of repair. Repeat the PCR from fresh lysates with a no-template control. Discard the plate's genotypes if the control fails.

If every clone is in-frame or unedited at a gene you expected to kill, the frameshift cells probably died. The plate is a museum of cells that escaped. That is biological information. It is not a set of knockouts with awkward sequencing.

If a clone looked pure and a later passage looks mixed, the well was double, the culture was cross-contaminated, or the edit is unstable because the nuclease is still present. Reclone from an early freeze. Do not average the two traces into a story about mosaicism you did not set out to study.

The cells that grow are a biased catalogue

Edited cells that hate the edit do not form colonies. Clones that expand are enriched for in-frame repairs, unedited alleles, and cells with extra unedited copies. If the phenotype you wanted was loss of an essential function, the clones on the plate are the ones that kept the function. Report the recovery rate. A stack of ten wild-type colonies is not ten failures of PCR. It may be selection.

Aneuploid lines hide a third copy. You can sequence two frameshifted alleles and leave one intact allele that supplies the protein. Copy number is part of the parental description, not an optional extra for difficult reviewers. Without it, homozygous is a hope.

Off-target and passenger mutations travel with the clone you like. Two independent clones with the same on-target allele and the same phenotype are stronger than one clone with a beautiful trace. If they disagree, the phenotype is not ready to be attributed to the target.

The parent hazard does not shrink

Picking a clone does not lower the biosafety category of the parent line. An edited human cell line remains under the approval that allowed the editing. This glossary does not authorise a new use, does not make the clone a diagnostic reagent, and does not describe embryo or germline work. Discard the pool and the failed wells under the same rules as the starting culture. A well that "looks empty" can still hold viable cells.

If the selection drug is toxic to people as well as to cells, follow the safety data for that substance. The molecular biology of resistance is not a handling instruction.

A weekend without power

Single-cell outgrowth is slow and easy to kill with a cold incubator. A power cut between plating and the first feeding can erase a plate or, worse, leave a few wells that grew because they were crowded doublets with more thermal mass of medium and a head start. After an interruption, do not rescue the surviving wells as your clonal series. Repeat the deposition. Humidity and loose plate seals change evaporation across a 96-well plate, so edge wells concentrate and bias which clones you meet. Rotate or note the edge effect rather than discovering it as a biological cluster of genotypes.

Freeze early passages once a genotype is real. A later contamination cannot be undone by resequencing a culture that is already mixed, unless an early vial exists. Freezer discipline for those vials is the same practical problem as storing biological samples from fridge to freezer.

What to ask for when colonies are not evidence

Tell the sourcing note how you will prove clonality and which alleles must be sequenced, including copy number if you know the line is not diploid. Name the selection drug only as a reagent class you need, not as the readout. Amplicon or clone sequencing can be discussed through the CRISPR validation sequencing enquiry reference. Treat that page as an enquiry reference. It is not a claim that clone validation is run as an operated service. The molecular biology pathway is the surrounding experimental frame. Ask whether a quotation is possible for the genotyping reagents and plasticware the plan uses. Do not describe a resistant colony as material that only needs confirmation.