A gel shift that "looks right" is not always a gel shift you can use. In an electrophoretic mobility shift assay (EMSA), a shifted band can arise from sequence-dependent binding, non-specific nucleic-acid affinity, heterogeneous complexes, or even aggregation. The difference between an informative experiment and an ambiguous one is usually control design—and the discipline to interpret the pattern based on what those controls actually test.
This resource focuses on how to design EMSA controls, interpret band patterns with appropriate caution, and report results in a way that supports internal review and downstream decisions. If you need the step-by-step execution workflow, see the protocol guide: Electrophoretic Mobility Shift Assay Protocol.
In EMSA, "specific" is best treated as a testable claim, not an impression. A practical definition is:
Equally important is what EMSA does not establish on its own. A single shifted band does not automatically tell you which factor is bound (especially in extracts), whether binding occurs in chromatin context, or how binding behaves kinetically. Those questions may require orthogonal methods, but within EMSA itself, the most reliable path is to build a control set that can support the claim you want to make.
Many "failed" EMSA projects are not failed experiments—they are under-controlled experiments. The goal is to build a minimum viable lane set that distinguishes sequence-dependent binding from background effects.
Minimum Controls and What Each One Proves
| Control Lane | What It Tests | What You Can Conclude (If It Behaves as Expected) |
| Free probe | Baseline migration and probe integrity | Probe is intact; free-probe reference is clear |
| Protein + probe | Complex formation under native conditions | A shifted species forms in the chosen buffer and run conditions |
| + Specific competitor (unlabeled, same sequence) | Sequence dependence | The shift reflects binding to that sequence element (not just "any nucleic acid") |
| Sequence-disrupted control (mutant probe or disrupted competitor) | Motif/site dependence | Binding depends on a defined site rather than probe length/composition alone |
| Optional: non-specific competitor | Background diagnosis | Lane improvements reflect reduced stickiness rather than loss of a true specific interaction |
| Optional: supershift | Identity support (context-dependent) | Antibody engagement is consistent with involvement of the target factor |
If you are using nuclear extracts or other complex mixtures, the "optional" controls become more important, because off-target binders and heterogeneous complexes are more likely.
Example EMSA lane set showing minimum controls to support specificity and clean interpretation.
Competition is one of the most informative gel shift assay controls because it tests the interaction under nearly identical lane conditions—only the competitor changes.
A specific competitor is an unlabeled oligo that matches the labeled probe sequence. If the complex truly depends on that sequence, the competitor should reduce the shifted signal by competing for binding.
Interpretation anchor:
If the shift decreases in a controlled way in the presence of a sequence-matched competitor, that supports sequence-dependent binding.
Common pitfalls
A non-specific competitor is an unrelated sequence used to suppress general nucleic-acid affinity or "stickiness." It is most helpful when your lanes show diffuse signal or high background.
Interpretation anchor:
If the non-specific competitor improves band sharpness and reduces background, it suggests the system had non-specific association that could obscure interpretation.
Use a simple logic chain rather than narrative interpretation:
Where competition tests "replaceability," mutant probes test "site dependence." A well-designed mutant probe is often the strongest way to support motif-dependent binding.
If you are building a broader protein–DNA binding workflow (multiple motifs, variants, or targets), coordinating probe logic and reporting across experiments becomes easier when EMSA sits inside a larger interaction strategy: Protein–DNA Interaction Analysis.
EMSA supershift is often described as an "identity check," but it should be treated as confirmatory, not definitive. In simple terms, you add an antibody and look for a mobility change consistent with antibody engagement.
Best practice: supershift should reinforce control-based specificity evidence. It should not replace it.
Interpretation is most defensible when you name the pattern and tie it to a control-supported explanation.
Controls-based decision tree for interpreting EMSA band patterns and identifying next steps.
Often consistent with a dominant complex. Specificity depends on how the band behaves with competition or sequence disruption—not its intensity alone.
Common explanations include heterogeneous binding states, oligomerization, multiple effective sites, or mixed complexes (especially in extracts). Use competition and mutant probes to determine which bands track with the intended site.
This pattern often indicates one of the following:
If smear is a persistent issue, address assay stability and run conditions before expanding the control set.
This can reflect weak binding, but can also reflect stickiness. Without sequence-based controls, it is hard to interpret reliably.
A practical rule: describe what you see, but only claim what your controls test.
Data quality in EMSA is not only a matter of signal strength. It is the combination of interpretability, repeatability, and sufficient documentation for review.
Quality Checks That Support Review and Comparison
| Category | What to Verify | Why It Matters |
| Lane completeness | Required controls are included | Prevents conclusions from under-defined lane logic |
| Free probe quality | Free probe is visible and not saturated | Preserves a reliable reference and usable dynamic range |
| Specificity evidence | Competition and/or mutant logic supports the claim | Anchors interpretation to an experimental test |
| Replicate behavior | Key lanes behave consistently across repeats | Builds confidence in the conclusion |
| Recording discipline | Key parameters are captured | Enables defensible comparisons across conditions |
Avoid "perfect gel bias." A visually strong gel with missing specificity controls is often less useful than a modest gel with correct lane logic.
A reviewer should be able to answer two questions: what exactly was run, and why the conclusion follows from the controls. A concise EMSA report typically includes:
Sometimes the fastest way to strengthen a conclusion is not another lane, but a method chosen for the uncertainty you still have.
This approach helps avoid over-fitting EMSA to questions it is not designed to answer.
Include free probe, protein + probe, and at least one sequence-based challenge (specific competitor or mutant probe). Add non-specific competitor or supershift only if they resolve a clear uncertainty.
A sequence-matched unlabeled competitor tests whether the shifted band depends on the probe sequence. A controlled reduction of the shift supports sequence-dependent binding under the same lane conditions.
It can indicate non-sequence-dependent binding, saturation, or an unsuitable competitor design. Re-check probe composition, lane consistency, and whether the competitor truly matches the binding element being tested.
Loss of the shift can occur if the antibody disrupts complex formation rather than stabilizing a larger complex. Treat this as supportive only when competitor/mutant controls already indicate sequence-dependent binding.
Use sequence disruption and competition to see which bands track with the intended binding site. Bands that do not respond to these controls are less likely to represent the target interaction.
The conclusion is not tied to what the controls test. Review-ready EMSA claims should explicitly reference competition or sequence-disruption outcomes, not band intensity alone.
Provide probe sequence/label, protein input description, buffer and competitor conditions, gel/run settings, a lane map with control rationale, raw images (non-saturated), and a conclusion linked to control outcomes.
Not by itself. Extract-based shifts can reflect multiple binders; identity support typically requires sequence-based controls plus confirmatory evidence (e.g., well-behaved supershift or an orthogonal method).
References
Online Inquiry