In situ hybridization
Single-molecule FISH tiles dozens of short, singly labelled oligonucleotides along one transcript, so a single mRNA becomes a diffraction-limited spot while any individual oligo bound elsewhere stays below the noise. The readout is a count of spots per cell, which makes the method genuinely quantitative.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
smFISH
Single-molecule FISH tiles dozens of short, singly labelled oligonucleotides along one transcript, so a single mRNA becomes a diffraction-limited spot while any individual oligo bound elsewhere stays below the noise. The readout is a count of spots per cell, which makes the method genuinely quantitative.
How it differs from the standard method
- Typically 30–48 oligos of about 20 nt; fewer oligos means a dimmer spot and a lower count at the same threshold.
- Needs a high-NA objective and a z-stack, because spots leave the focal plane within about a micrometre.
- Spot calling needs an explicit threshold, checked by sweeping it and sitting on the plateau.
- Works best on cultured cells and thin sections; thick tissue and autofluorescence quickly become limiting.
1A labelled probe finds its transcript where it lies
In situ hybridization keeps the sample intact and brings the detection reagent to the RNA, instead of the other way round. The cells or the section are fixed so that the transcripts stay where they were, opened up so a nucleic-acid probe can diffuse in, and then incubated with a probe complementary to the target.
2Specificity is designed into the probe and enforced by the wash
Probe
Length, GC content and melting temperature set what the probe will tolerate. Short oligos tiled across the transcript give specificity from the requirement that many of them bind the same molecule.
Stringency
Hybridisation and wash conditions set how close to perfect a duplex has to be to survive. Low stringency keeps mismatched duplexes; high stringency melts the matched one too.
Control probe
A probe with the same label, length and GC content that cannot hybridise to the target. Every spot it gives is background, and that number is your false-positive rate.
Autofluorescence
Lipofuscin, elastin, collagen and red blood cells fluoresce. Lipofuscin granules in particular are punctate, dot-sized and bright in every channel.
3Fixation is a trade-off in both directions
Cross-linking fixation has to hold the RNA in place and keep the morphology you came to look at, while still letting a probe through. Under-fixation loses the RNA: tissue RNases keep working until the formalin reaches them, and in cultured cells the cytoplasm leaks before the cross-links form.
4From judging intensity to counting molecules
Classical ISH gives a stain you describe as strong or weak. Two developments made it quantitative.
5How it fails
- Spots on the glass as well as in the cells: stringency too low, so the probe is holding on to partial matches.
- A count far below a matched RT-qPCR number: probe access blocked by over-fixation, too few oligos in the set, or a threshold set past the plateau.
- Bright punctate specks in the negative-control section: tissue autofluorescence, usually lipofuscin, not signal.
- Beautiful morphology and no signal anywhere, control probe included: the RNA degraded before or during the run.
- A count that changes a lot when you nudge the threshold: you are on the slope, not the plateau, and the number is not yet a measurement.
Key terms
Practise these as flashcards →How close to perfectly complementary a duplex must be to survive hybridisation and washing; set by formamide, salt and temperature.
A control probe with the same label and physical properties that cannot hybridise to the target; its signal is background by definition.
Deciding which intensity maxima in an image are single molecules, usually by a threshold expressed in multiples of the local background standard deviation.
Heating a fixed section in buffer to reverse part of the formalin cross-linking so a probe can reach the RNA.
Two half-probes that must hybridise adjacently before an amplification cascade can assemble, so a single mis-bound probe gives no signal.
Autofluorescent granules that accumulate in long-lived cells; punctate, dot-sized and bright in every channel, so they mimic real spots.
The smallest spot an optical system can form; single molecules appear at this size, so resolving them needs a high-numerical-aperture objective.
When to use it
- You need to know which cells in a mixed population express a transcript, not the population average.
- You need the position of a transcript inside a cell: nuclear versus cytoplasmic, polarised, in a process.
- You have an expression difference from RNA-seq or RT-qPCR and need to see which cell type in the tissue it comes from.
- You want an absolute count of molecules per cell, with a threshold and a control you can defend.
- Only fixed or archival material exists, and a lysate method would destroy the information you need.
Compared with alternatives
Far more sensitive and quantitative for how much, across a whole sample, but reports nothing about which cell or where in it.
Prefer it when you do not already know which genes to probe; it maps many genes at once instead of one or a handful.
Shows the protein rather than the transcript; use it when the antibody is good, and ISH when it is not or when you care about transcription itself.
Gives whole transcriptomes per cell, but dissociation throws the tissue coordinates away; ISH is how such results are confirmed in situ.
Use it to confirm that a probe recognises a transcript of the expected size and splice form before trusting spots in a cell.
Sub-techniques
Choosing one here switches the whole page, including the protocol and the lab, to it.
Single-molecule FISH tiles dozens of short, singly labelled oligonucleotides along one transcript, so a single mRNA becomes a diffraction-limited spot while any individual oligo bound elsewhere stays below the noise. The readout is a count of spots per cell, which makes the method genuinely quantitative.
- Typically 30–48 oligos of about 20 nt; fewer oligos means a dimmer spot and a lower count at the same threshold.
- Needs a high-NA objective and a z-stack, because spots leave the focal plane within about a micrometre.
- Spot calling needs an explicit threshold, checked by sweeping it and sitting on the plateau.
- Works best on cultured cells and thin sections; thick tissue and autofluorescence quickly become limiting.