Almost every CRISPR experiment starts with a decision about format rather than about biology: everything on one plasmid or the nuclease and the guide apart, a plasmid or a virus, transient or selected, your own guide cloned in or a catalogue construct. That decision sets how long the work takes and how believable the result is, because a format that delivers poorly into your cells produces an edited minority and a phenotype that is diluted by whatever did not take. Choosing it deliberately is cheaper than optimising a transfection for a construct that was never going to fit.
Which crispr cas9 plasmid format to order
An all in one vector carrying the nuclease, the guide cassette and a marker is the fewest steps and the largest plasmid, which matters because transfection efficiency falls with size and the difference is not small in primary cells. Separate vectors transfect better and need both to arrive in the same cell, which is fine in a line that takes DNA well and poor in one that does not. Where the cells are hard to transfect, the answer is usually neither: a ribonucleoprotein or a viral delivery moves the problem out of the plasmid entirely, and a plasmid experiment that has failed twice in such cells rarely succeeds on the third attempt.
What a crispr kit contains, and which crispr cas9 products it replaces
A kit is a validated pair plus the reagents to get it into cells and, usually, a control guide and a way of checking the edit. What it replaces is the assembly work: cloning a guide, sequencing the construct and proving the pair cuts. That is worth paying for on a first experiment and worth doing yourself once the laboratory edits regularly, because the per experiment cost of a kit is several times the cost of the parts. Compare kits on the delivery format and on whether the guide sequences are published, because an unpublished guide cannot be analysed for off target sites.
Where crispri differs from a knockout, and what it needs from the plasmid
Interference silences transcription instead of cutting: a catalytically dead nuclease fused to a repressor sits at the promoter and turns the gene down while the sequence stays intact. That makes it reversible, titratable and usable on an essential gene, which is the case a knockout cannot answer, and it makes the plasmid different, because the fusion is larger and the guide has to target a window near the transcription start rather than a coding exon. The readout also changes: there is no indel to sequence, so the evidence is the transcript and the protein rather than the allele.
Plasmid libraries and crispr screening
A screen is the same chemistry at a different scale, and the plasmid becomes a library of thousands of guides delivered so that most cells receive one. Everything then depends on representation: how many cells per guide were infected, how many were kept at every passage, and whether the selection window is long enough for a real effect and short enough that drift does not dominate. The library is the cheap part of a screen and the coverage is the expensive part, which is why a screen designed around the number of cells that can actually be handled is the one that produces a readable result.
Cloning your own guide against buying the construct
Cloning a guide into a published backbone is an afternoon and a sequencing reaction, and it is the right answer when several guides per gene are needed, which is most of the time. Buying the construct is the right answer when the guide is standard, when the backbone is unusual, or when the time matters more than the money. Whichever it is, keep the guide sequence, the backbone name and the verified plasmid map in the record, because the next person to repeat the experiment needs all three and none of them is recoverable from a catalogue number.
What to check before the plasmid goes near cells
Three things, and all are cheap. Sequence the guide cassette rather than trusting a ligation, because a wrong or truncated guide looks exactly like a failed experiment. Confirm the plasmid preparation is endotoxin low if the cells are primary, since that alone can explain a poor transfection. And run a cutting test in an easy line first, so that a negative result in the cells you care about can be attributed to delivery rather than to the pair.
cas9 plasmids and the crispr vector around them
A CRISPR vector is a delivery decision as much as a construct. Cas9 plasmids come in three shapes: all in one, carrying nuclease and guide together, which is the simplest for a single edit; nuclease only, for a cell line that will receive many guides; and guide only, for a line that already expresses Cas9, which is what a pooled screen needs. Choose the promoter for the cell, not for the brightness of the data, and check the nuclease variant, since a high fidelity mutant, a nickase or a base editor change what the plasmid is for entirely. The backbone then decides delivery: transfection, lentivirus or in vitro transcription.
Questions people ask about crispr plasmids
All in one or separate vectors?
All in one for a line that transfects well, separate vectors when efficiency is the constraint and both can be delivered at high multiplicity, and neither when the cells are genuinely hard to transfect. In that case a ribonucleoprotein or a virus is the honest answer rather than another round of plasmid optimisation.
Do I need to sequence the construct?
Always the guide cassette, and ideally the junctions around it. A wrong guide is indistinguishable from a failed edit at the point where you are trying to interpret a phenotype, and the sequencing costs a fraction of the experiment it protects.
Is interference a substitute for a knockout?
For an essential gene and for anything where reversibility or a dose response matters, it answers questions a knockout cannot. For a clean statement that a protein is absent it is not, because silencing is partial by nature. Many programmes use interference to find the genes and a knockout to make the statement.