1. Define exactly what you need to detect

Identify the target using its gene/protein name and an accession identifier where possible. Decide whether you need total protein, a particular isoform, a cleavage product or a specific post-translational modification. For a phospho-specific antibody, specify the residue and relevant sequence context.

Check the immunogen and mapped epitope, when disclosed, against the protein form present in your model. An antibody recognising a shared region may detect several isoforms; a region absent from a truncated protein cannot report that form. An immunogen description alone does not demonstrate specificity in a complex sample. [4] [7]

Selection check: distinguish what the antibody is advertised to recognise from what the experimental data actually establish.

2. Separate target-species reactivity from antibody host species

The target species describes the origin of the antigen; the host species describes the organism in which an antibody was originally raised. An antibody raised in rabbit may recognise a human protein. Check tested reactivity with your target species and distinguish experimental evidence from sequence-based predictions.

In transfected models, specify both the introduced protein and the background species. For indirect detection in tissue, consider endogenous immunoglobulins: using a mouse primary on mouse tissue, for example, can produce background when an anti-mouse secondary also detects tissue immunoglobulins. Alternative host species or an appropriate same-species staining strategy may be needed. [4]

Selection check: assess target recognition and compatibility with the detection system as two separate decisions.

3. Evaluate application-specific evidence, not just a validation label

Inspect the actual images, controls, sample identity and protocol supporting the reference. A product validated in Western blot is not thereby validated in IHC or flow cytometry; even within one application, performance depends on sample context and assay conditions. [1] [2]

Prefer evidence showing endogenous target detection in a relevant sample, supported by an appropriate specificity strategy. Purified antigen or overexpression data can be informative, but alone may not establish selective detection at endogenous expression levels. Published use is useful when the methods and controls match your needs; citation count alone is insufficient. [4] [7]

Selection check: ask which evidence supports the exact clone or preparation, species and application you intend to use.

4. Match epitope accessibility to sample preparation

Preparation can mask, expose or alter epitopes. Compare your protocol with the conditions used to characterise the antibody rather than relying on the application abbreviation alone. [5]

IHC: distinguish FFPE from frozen tissue and compare fixative, fixation conditions and antigen retrieval. Western blot: check compatibility with denaturing versus native conditions, heating and reduction. [4] [7] Flow cytometry: live-cell surface staining requires an accessible extracellular epitope. [8] Conventional intracellular staining generally uses fixation and permeabilisation; select a clone and conjugate compatible with those conditions. [8]

Selection check: look for data obtained with a comparable preparation; successful use on one sample format does not establish suitability for every other format.

5. Choose the clone and production format for your experiment

A monoclonal preparation derives from one antibody clone; polyclonal preparations contain antibodies against multiple epitopes. Polyclonals can provide useful multi-epitope detection, but animal-derived preparations may vary between batches. Neither monoclonality nor a strong signal guarantees freedom from off-target binding. [4]

Recombinant describes production from antibody-encoding sequences, not a separate clonality category. Sequence-defined recombinant reagents support renewal and engineering of a defined antibody, but still require assay-specific characterisation. Consider the exact clone and conjugated format, continuity of supply and evidence for consistency between lots. [6]

Selection check: prioritise relevant performance data over a blanket preference for monoclonal, polyclonal or recombinant products.

6. Select the detection format as part of the complete assay

For indirect detection, confirm that the secondary recognises the primary’s host species and immunoglobulin class/subclass as required. In multiplex staining, check cross-reactivity between detection reagents and the other primaries, using appropriately designed omission controls. [3] If labelling the primary yourself, verify that its formulation is compatible with the chosen conjugation chemistry. [4]

In flow cytometry, choose the clone–fluorophore combination with the instrument, antigen expression and other panel markers in mind. Spectral overlap and spreading affect resolution; brightness alone is not a sufficient selection rule. Confirm compatibility with fixation/permeabilisation. Use appropriate single-stain controls for compensation or spectral unmixing, and FMO controls where useful for gate placement. [8]

Selection check: evidence for an unconjugated antibody does not replace assessment of the conjugate in the final panel.

7. Plan complementary validation, controls and titration

The five-pillar framework was proposed by Uhlén et al. and subsequently adapted by Edfors et al. for Western blot. These are complementary options, not five mandatory tests for every antibody. Select feasible strategies that answer the relevant specificity question in your assay. [1] [2] [10]

• Genetic strategies: compare signal after target knockout or knockdown. Limit: incomplete depletion, residual isoforms or biological effects of gene disruption can complicate interpretation.
• Orthogonal strategies: compare the antibody readout with an antibody-independent measurement across suitable samples. Limit: correlation alone may be misleading; mRNA abundance is not interchangeable with protein abundance.
• Independent antibodies: compare antibodies recognising different, non-overlapping epitopes. Limit: agreement is not conclusive if the reagents share off-target reactivity or have poorly characterised epitopes.
• Tagged-protein / recombinant-expression strategies: assess recognition of an independently identified expressed target. Limit: tags and overexpression can alter abundance, localisation or epitope accessibility and may not represent endogenous protein.
• Immunocapture followed by mass spectrometry: identify proteins enriched by the antibody. Limit: associated proteins can co-purify; capture performance does not establish specificity in fixed tissue or on a blot.

Edfors’ Western-blot adaptation includes gel-based capture-MS analysis of proteins migrating at the detected position. This should not be conflated with immunoprecipitation-MS, and co-migration or limited MS sensitivity can restrict interpretation. Validation remains application- and context-specific. [1] [2] [10]

Include target-positive and target-negative samples. For Western blot, expected band mass alone is not proof of specificity; additional bands may reflect isoforms, processing or modifications. [7] For IHC/IF, no-primary controls test secondary/detection background, not the specificity of the omitted primary. [3] For flow cytometry, FMO controls support gate placement but do not establish target specificity. [8]

Titrate the reagent in representative samples rather than treating a datasheet dilution as universal. Optimise useful separation with acceptable background, not simply the strongest signal. Record the supplier, catalogue number, clone, lot, conjugate, RRID when available and final protocol; reassess after relevant reagent or protocol changes. [9] [7]

Selection check: define both the evidence a validation strategy can provide and the uncertainty it leaves unresolved.