Controls and Validation for Western Blot
Every Western blot needs the same set of controls, but not every control is needed for every experiment. This is the practical checklist — positive, negative, loading, no-primary, no-secondary — with a decision tree for when each is strictly required and when it is optional.
A Western blot control is a lane whose expected outcome is known independently of the experimental question. If the control lane behaves as predicted, the experimental lanes can be interpreted; if it does not, the experiment tells you nothing about the biological question and everything about the assay. Skipping controls to save a lane is the single most reliable way to generate unpublishable data.
This protocol covers the five standard control types, when each is strictly required, and how to decide which subset your specific experiment needs. It complements the more detailed protocols on dilution optimization and troubleshooting non-specific bands.
The five standard control lanes
1. Positive control
A sample known to contain the target at detectable levels. Demonstrates that the antibody detects the target under your specific conditions. If the positive control is blank, no conclusion can be drawn from the experimental lanes — the assay is not working, and any signal or lack of signal in the sample lanes is uninterpretable.
2. Negative control
A sample known not to contain the target. Demonstrates that the antibody signal in the experimental lanes is target-specific, not background or cross-reactivity. If the negative control gives signal at the target MW, the antibody is not specific enough for your sample type, or the sample preparation is contaminating the negative material.
3. Loading control
A housekeeping protein probed on the same blot (or a total-protein stain) that reports lane-to-lane loading consistency. Enables comparison of experimental band intensity across lanes. Without a loading control, apparent "upregulation" of the target may simply be more protein loaded in that lane.
4. No-primary control
A lane loaded with sample but incubated with buffer instead of primary antibody, then processed normally through secondary and detection. Demonstrates that any signal in the blot requires primary antibody — ruling out non-specific secondary binding as the source. If the no-primary lane has signal, your secondary is binding directly to something in the sample.
5. No-secondary control
A lane loaded with sample and incubated with primary but not secondary, then processed through detection. Demonstrates that any signal requires the secondary + detection chemistry — ruling out primary autofluorescence or ECL substrate reactivity with the primary alone. Rarely needed but essential when the primary is directly conjugated (HRP or fluorophore) and you want to distinguish direct conjugate signal from indirect.
The positive control — what counts as "known to contain the target"
A positive control lysate needs to satisfy three criteria: the target is expressed at detectable levels; the expression level is stable and reproducible across preparations; and the sample is available in sufficient quantity to serve as a standard across multiple blots.
In practice, positive controls fall into three tiers:
Tier 1 — recombinant protein spike-in
Purified recombinant target protein loaded at defined concentration (10-100 ng typically). This is the gold-standard positive control because the identity, purity, and quantity are all defined. It runs slightly differently from endogenous protein (no PTMs, sometimes tagged) but establishes the exact MW your antibody should detect. Triple Point supplies recombinant proteins for a subset of targets — e.g., recombinant BACE1, recombinant BACE2 — specifically for this use.
Tier 2 — validated cell line or tissue lysate
A lysate from a cell line or tissue known to express the target at moderate to high levels. HEK293 cells expressing the target (endogenous or overexpression), primary cell types with published expression (e.g., macrophages for cathepsin B/L, neurons for BACE1), or tissue samples with well-characterised expression (liver for cathepsin D, brain for BACE1). The MW of the endogenous form matches the sample, which is more useful than recombinant when interpreting PTMs.
Tier 3 — overexpression system
Transient transfection of the target-encoding plasmid into an easy-to-transfect cell line (HEK293T, COS-7). Higher expression than endogenous, useful when the target is expressed at low levels and you need a bright signal for antibody characterisation. Not appropriate as a quantitative reference — overexpression levels are transfection-dependent and can skew perception of "normal" MW.
The negative control — how strong a "no signal" claim you need
Negative controls range from convenient-but-weak to definitive-but-costly:
KO cell line (strongest)
Cells where the target gene has been genetically deleted by CRISPR or other knockout technology. If the antibody gives no signal on a KO lysate, it is truly target-specific — no other protein in the cell binds it at detectable levels. This is the definitive negative control. See our CRISPR knockout protocol for setting one up.
siRNA knockdown (strong)
Cells treated with siRNA against the target, showing 60-90% reduction of protein. If the antibody signal is proportionally reduced, it is target-specific. Not quite as clean as KO because residual target protein remains, but faster and more accessible for most labs.
Peptide competition (moderate)
The primary antibody is pre-incubated with an excess of the immunogen peptide before use on the blot. If the signal is abolished, the antibody is specific for the peptide it was raised against. Note that this does not prove specificity for the intact protein containing that peptide — only for the epitope. Useful as a fast validation for a polyclonal antibody where the immunogen peptide is available.
Non-expressing tissue or cell line (moderate)
A sample known from independent evidence (published expression atlases, RNA-seq, mass spectrometry) not to express the target. Weaker than KO because the sample may contain low-level expression that is masked by other assays. Adequate for routine controls when KO is not available.
Species-mismatched sample (weakest, use with caution)
A sample from a species where the target antibody is not expected to cross-react. Only valid if orthologue reactivity is documented on the CoA. Do not assume species mismatch means no cross-reactivity — many antibodies against conserved regions detect orthologues from unrelated species.
Loading controls — house keepers vs total protein
Loading controls fall into two categories: individual housekeeping proteins, and total-protein stains.
Housekeeping proteins
β-actin (42 kDa), GAPDH (37 kDa), α-tubulin (55 kDa), β-tubulin (55 kDa), HSP90 (90 kDa), vinculin (117 kDa). Common, cheap, well-characterised. Assumes stable expression under experimental conditions — not always true. GAPDH varies with metabolic stress; β-actin varies during cytoskeletal reorganisation; tubulins vary in dividing cells.
Total protein stains
Ponceau S (pre-block), Amido Black, or Revert / No-Stain Protein Labelling (post-transfer). Reports the total protein per lane without dependence on any specific housekeeping protein's stability. Recommended by the recent Western blot reproducibility literature as the more robust approach for quantitative comparisons. Preferred for publication-grade quantitation.
Whichever you choose, the loading control lane must be run on the same blot as the target — a separately-run duplicate blot does not control for transfer efficiency or membrane variability.
A decision tree — which controls does my specific experiment need?
The five controls are the full catalogue. For any given experiment, some subset is strictly required and the rest are optional. Use the following decision tree:
Are you validating a new antibody or new sample type for the first time?
Required: all five. Positive, negative, loading, no-primary, no-secondary. You do not know how the antibody behaves on this sample yet; every control provides orthogonal information.
Are you running a routine experiment with a well-characterised antibody + well-characterised sample type?
Required: positive control, loading control. Optional but recommended: negative control every third or fourth experiment as a spot-check. No-primary and no-secondary are typically omitted after the initial characterisation.
Are you making a quantitative claim about a change in expression?
Required: positive control, negative control, loading control (preferably total-protein stain). Quantitative comparison requires all three. Additionally: run technical replicates within the same blot (duplicate lanes) and biological replicates (independent samples) to establish variance.
Are you claiming target specificity of the antibody in a new publication?
Required: positive control (recombinant or high-expression), negative control (KO preferred, siRNA acceptable), no-primary control, and either two orthogonal antibodies against non-overlapping epitopes or a peptide competition. This is the IWGAV (International Working Group for Antibody Validation) standard for publication-grade antibody validation.
Are you troubleshooting an experiment that failed?
Required: whichever control corresponds to the failure mode you suspect. Weak signal — positive control confirms the assay works. Non-specific bands — negative control isolates specificity. High background — no-primary and no-secondary controls isolate the source. See weak signal troubleshooting and non-specific band troubleshooting for structured flows.
Common control failure modes and interpretation
Positive control blank
The assay is not working. Sources: primary antibody degraded (check age, storage), primary dilution too weak, secondary antibody degraded, transfer failure, ECL substrate expired, sample lost or not loaded. Rerun the assay before drawing any conclusions from the experimental lanes.
Negative control gives signal at the target MW
The antibody is not target-specific in this sample type — it detects something else at the same MW. Options: change to a different antibody, change to a paralogue-specific antibody, use a stronger negative control (KO instead of siRNA), or run a peptide competition to confirm the observed signal is at least the correct epitope.
Loading control uneven across lanes
Loading is not consistent. If the variation is <20%, quantitation with loading-control normalisation is defensible. If >20%, rerun with more careful sample loading. Consider using a total-protein stain instead of a housekeeping protein, which is more robust to variation in specific housekeeper expression.
No-primary control gives signal
The secondary antibody is binding non-specifically to something in the sample. Sources: secondary too concentrated, sample contains endogenous IgG (typical of tissue lysate) that secondary is detecting, blocking insufficient. Solutions: reduce secondary concentration 5-10×, use pre-adsorbed secondary for cross-species blots, extend blocking time or switch blocking buffer.
No-secondary control gives signal
If using a directly-conjugated primary (HRP or fluorophore), any signal here is direct-conjugate signal — expected and interpretable. If using an unconjugated primary, no-secondary signal indicates ECL substrate is reacting with something else on the membrane, typically endogenous peroxidase activity. Rare but occurs with tissue lysate (red blood cell contamination). Quench with 0.3% H2O2 before probing.
Control lanes and blot layout
Layout matters. Positive control should be on the far left (adjacent to the ladder) and the negative control on the far right, so the eye can compare experimental lanes against both extremes without scanning across the whole blot. Loading control probing happens on the same membrane after target-specific probing, either by stripping and reprobing or by cutting the membrane and probing separately at different MW regions.
If the blot is large enough (12-15 lanes), include two positive-control lanes at different loading amounts (10 μg and 30 μg) to establish the linear range for quantitative comparison. If it is small (6-8 lanes), one positive and one negative is the compromise.
Documenting controls in the lab notebook
Every blot's control lanes should be recorded in the same way, so cross-experiment comparison is possible. Record: which cell line or tissue was the positive control, its passage or preparation date, the target expression level (if known from prior characterisation), which cell line or intervention was the negative control, the housekeeping protein or total-protein-stain used, and any deviation from the standard control panel. Over time this record becomes the reference for what "normal control behaviour" looks like in your lab, and unusual results become immediately obvious.