01Define which distinction the assay must make

ICH Q2(R2) describes specificity/selectivity through absence of interference, comparison with an orthogonal procedure or justified inherent discrimination. Where one procedure is insufficient, complementary procedures can supply the necessary distinction.ICH — Validation of analytical procedures Q2(R2) (opens in a new tab)

The required distinction depends on the task. An identity test must discriminate the target from relevant alternatives. A content assay must quantify the target without materially counting another component as part of it.

A method can be sufficiently selective for one purpose without resolving every molecule in a mixture. The claim should name that purpose rather than imply universal chemical identification.

02Do not confuse several readings with independent evidence

In targeted mass spectrometry, a precursor-to-product transition is a selected signal channel. Monitoring several channels can help assess consistency, but each still needs to be examined for interference.

Abbatiello and colleagues developed AuDIT to identify problematic transitions in peptide quantification. Their primary study compared relative product-ion patterns with isotope-labelled standards and considered replicate variability, addressing interference that could otherwise distort quantitative results.Abbatiello and colleagues — Detecting problematic transitions in peptide quantification (opens in a new tab)

The finding is not that a particular number of transitions guarantees specificity. It is that signal agreement and inconsistency can be investigated rather than assuming that every selected channel is clean.

A standard’s expected response is a useful comparison, but the surrounding matrix and instrument conditions must remain relevant to that comparison.

03See how co-measurement can produce a plausible total

For an original simplified example, suppose the intended peptide contributes 90 response units and an unresolved impurity contributes 10. If the assay interprets all 100 units as target peptide under an equal-response assumption, it overstates the target contribution.

The observed total can be highly repeatable. Repeating the same unresolved measurement does not separate the two contributions or correct the bias.

Adding a known amount of target may also produce the expected increase while the original 10-unit contribution remains. A successful incremental response is not, on its own, proof that the unspiked signal was free of interference.

Real components need not have equal response factors. The simplified example demonstrates the logic of co-measurement; it does not provide a correction formula for a real chromatogram.

To resolve the ambiguity, the additional evidence must address the competing contribution: through suitable separation, a discriminating signal or another justified approach.

04Keep the challenge set and remaining limits visible

A useful specificity assessment states which plausible interferences were investigated and why. Testing an unrelated clean solvent cannot substitute for examining a relevant close analogue or degradation product.

If an interference standard is unavailable, that limitation should remain explicit. Suitable characterised mixtures or an orthogonal comparison may provide evidence, but the conclusion must match the alternatives actually addressed.

A new sample matrix or changed impurity profile can challenge a previously suitable assay. A validation date is therefore not a permanent guarantee independent of the material and method conditions.

Report the supported claim precisely: the procedure discriminated the target from the tested alternatives at the relevant levels. That tells the reader more than a blanket assertion that the instrument is highly specific.