A knockout is a claim about a protein, and most of the work is proving it rather than making it. Editing is now the easy part: a guide, a nuclease and a week of culture will give a pool with indels at the right locus. What takes the time is isolating clones, showing that every allele is disrupted in a way that cannot produce a functional product, and showing that what you see afterwards is the loss of that protein rather than the clone being a clone. A line bought or built without that evidence is a line whose phenotype will be argued about later.
Making a knockout cell line, and what a knockout cell on its own is not
A pool of edited cells is not a knockout: it is a mixture whose average is reduced, and any phenotype measured in it is diluted by whatever fraction escaped editing. Single cell cloning is what turns the pool into a line, and it is also what introduces the problem that makes this work slow, because a clone differs from its parent in more than the edit. The defence is several independent clones and, where the phenotype matters, a rescue that puts the protein back.
Reading a ko cell line certificate
Three things belong on it and often only the first is there. The sequence of every allele at the target site, so a reader can see that each is disrupted and how; evidence at the protein level with an antibody validated against the region the edit does not remove, because a truncated product can be invisible to one antibody and active in the cell; and the identity and mycoplasma status of the clone. Where a supplier provides a chromatogram of a pool rather than allele sequences of the clone, the certificate is describing a different thing from the line.
Where a knock in cell line is harder, and why a knockin cell line takes longer
Adding a sequence depends on homology directed repair, which competes with the error prone pathway and loses in most cell types, so the yield of correct events is a small fraction of the edited ones. That means more clones screened, a screen that has to distinguish a correct insertion from a partial one, and usually a selection cassette whose removal is its own step. Budget it as a project with a screening plan rather than as a knockout with an extra template.
When a cas9 stable cell line is the right host
Holding the nuclease in the cell makes sense when you will edit the same background repeatedly: only the guide has to be delivered, which is cheaper, faster and more efficient in cells that transfect poorly. The price is a line that carries a foreign nuclease constitutively, with the off target exposure and the immunogenicity that implies, so it is the wrong host for anything that will be studied for its own biology or move towards a therapeutic use. For a screening background it is usually the right answer.
What a crispr knockout kit leaves you to do
A kit supplies the guide and the nuclease in a deliverable form and validates that the pair cuts. Everything after that is yours: the delivery into your cell type, the cloning, the allele sequencing, the protein level evidence and the controls. That is not a criticism of kits, it is what they are, and the useful question when comparing them is which delivery format suits your cells and whether the guide sequences are published so that an off target analysis is possible.
Where crispr gene knockout stops being the answer
Three places. An essential gene, where the cells that survive are the ones that did not lose it, and the honest tool is a degron or an inducible knockdown. A gene with redundant paralogues, where a single knockout shows nothing and the experiment is a double or a triple. And a gene whose transcript is the point rather than its protein, where an edit that disrupts translation leaves the RNA doing whatever it was doing. Choosing the tool by the question rather than by availability saves a year.
Questions people ask about knockout cell lines
How many clones do I need?
At least two or three independent ones for any phenotype you intend to publish, and a rescue if the phenotype is the point. One clone supports a statement about that clone; several supports a statement about the gene.
Is a frameshift enough evidence?
Not on its own. Frameshifts can produce a shortened product with residual activity, and cells sometimes initiate downstream of the edit. Pair the allele sequences with a protein level read using an antibody that would still see a truncated product.
Should I buy the line or build it?
Buy it where the line is a standard background and the supplier's evidence is complete, because their screening has already been paid for. Build it where the background is yours, where the edit has to be precise, or where you will need several independent clones of the same edit.