This tool covers antibodies targeting proteins and post-translational modifications. If you are using antibodies against non-protein targets (DNA, lipids, carbohydrates), the core principles still apply but the specific control strategies will differ. You can still use this tool to document your plan.
Evidence
Controls
Selection
Before you start
Not every antibody needs the same validation
This tool helps you plan proportionate validation for every antibody in your project. The level of scrutiny depends on what your scientific question demands from each antibody.
Over half of commercial antibodies fail rigorous independent testing, and most published papers using poorly performing antibodies present no relevant validation data. Planning before you start prevents wasted time, samples, and funding.
You'll categorise each antibody into one of three groups
Target-specific → full validation plan
Your scientific question is about this specific protein or modification. If the antibody is detecting something else, your conclusion is wrong.
e.g. "We hypothesise this protein is upregulated in human lung tissue in disease" — the antibody must be detecting that protein specifically.
Community marker → use the right clone
The antibody identifies a cell population or phenotype — it defines the context of your experiment, not the subject of it.
e.g. "I want to identify T helper cells the same way other immunology labs do" — use the HCDM workshop-verified clone, not just any anti-CD4.
Caution: markers for rare or less-studied populations may have a thinner evidence base than assumed.
Technical function → document the purpose
The antibody serves a technical role where specificity for the stated target is not what matters. Document what it's actually for.
e.g. anti-GAPDH as a Western blot loading control. Tip: total protein stains (Ponceau S, Stain-Free) are often more reliable.
Key principles
1
Match controls to your experiment. Off-target binding depends on what else is expressed in your sample. Controls in a different cell type or application are a starting point, not proof.
2
Search before you plan. Existing characterisation data (e.g. from YCharOS/OGA) can tell you which antibodies are likely to work before you commit to expensive validation.
3
Validate in the exact assay you're using. An antibody validated for Western blot is not validated for immunofluorescence — the protein is presented differently.
4
Prioritise recombinant antibodies. Independent testing shows ~67% WB success rate vs ~41% (monoclonal) and ~27% (polyclonal).
No OGA characterisation data for this target yet. You can still plan your validation below.
Community markers should be monoclonal or recombinant monoclonal antibodies. Polyclonal antibodies have too much lot-to-lot variability to serve as standardised tools. Specify which clone you plan to use.
Existing evidence
What independent characterisation data already exists for this target?
Why this matters: Independent testing by the YCharOS consortium found that over 50% of 614 commercial antibodies failed characterisation in at least one application. Pass rates vary: ~50% for Western blot, ~44% for IP, ~36% for IF. Recombinant antibodies consistently outperform polyclonal (~67% vs ~27% WB success). 84% of papers using failed antibodies presented no validation data at all.
Searching for existing characterisation data before you start can prevent months of wasted work on an antibody that has already been shown to fail.
PTM-specific antibodies have an additional validation dimension. A genetic knockout removes the entire protein — it confirms the antibody binds something on the right protein, but not that it distinguishes the modified from the unmodified form. Step 2 will guide you through modification-specific controls.
Workshop-verified clones for CD markers (flow cytometry)
Vendor datasheets
Check for KO-controlled data — replicable >80% of the time
What to look for
Prioritise:Recombinant antibodies (~67% WB success vs ~27% polyclonal)
Check:Same application and sample type as your experiment
Watch for:Tick-box claims without images — look at the actual data
Control strategy
Controls should match your experimental context. Off-target binding depends on what else is expressed in your sample.
Aim for genetic controls (knockout or knockdown) in the application you're using. If genetic controls are not feasible (essential gene, primary human tissue), the IWGAV framework provides alternative approaches: orthogonal methods, independent antibodies, recombinant expression, capture mass spectrometry. These are not fallbacks — document which approach you will use and why.
PTM-specific controls needed. A knockout removes the entire protein — it doesn't confirm modification specificity. You also need: enzymatic removal (e.g. lambda phosphatase), stimulation/inhibition conditions that modulate the modification, or blocking peptide comparisons (modified vs. unmodified).
Best:Wild-type from a KO pair (verify the KO independently — ~30% of commercial KOs may not be true knockouts)
Good:CRISPR knock-in at endogenous levels, with or without tag
Good:Lentiviral stable expression with controllable expression level
Quick screen:Commercial overexpression lysate (~£200) — shows expected MW on blot
Gold standard:Genetic KO in your cell type (verify independently)
Good:KO in a different cell type — the YCharOS/OGA approach for initial antibody selection
When KO not feasible:siRNA/shRNA knockdown (confirm efficiency by RT-qPCR; beware siRNA off-target effects)
Weaker:Cell line/tissue without target expression (check proteomic/transcriptomic datasets)
For IHC on human tissue: consider staining FFPE cell pellets from KO lines alongside your tissue sections, using the same protocol.
A control that isn't run alongside your experiment can't account for day-to-day variation. A control in a different cell type or tissue can't account for differences in off-target binding — the proteome is different, so the opportunities for cross-reactivity are different. Document these gaps honestly: a clean blot in HEK293 doesn't guarantee a clean result in primary tissue. For flow cytometry, fixation and permeabilisation can fundamentally change antibody performance.
If your antibody shows increased protein expression, does RT-qPCR show increased mRNA? If flow cytometry shows a shift, does single-cell RNA sequencing support the same conclusion? These complementary approaches strengthen any antibody-based finding. They do not replace validation — they corroborate it.
Antibody selection
At planning stage you may not have chosen a specific antibody yet. Describe what you'll look for — the full product identity belongs in the Validation Recorder once you have results.
Where possible, test at least two independent antibodies targeting different epitopes. If both give the same result, confidence in specificity is much higher. This is one of the IWGAV five-pillar strategies and is especially valuable when genetic controls are not feasible.
