What do repeatability and intermediate precision mean for peptide methods?
Precision is the closeness of agreement among independent results obtained under stipulated conditions. It is distinct from trueness, specificity, linearity and robustness. In peptide analysis, weighing, dissolution, dilution, transfer, chromatography and data processing can each contribute to variability.
Repeatability evaluates precision over a short interval with the same procedure, analyst and equipment. The scope must be explicit. Repeated injections from one vial assess injection-level repeatability; independently prepared portions of the same homogeneous sample can assess preparation-inclusive method repeatability. Intermediate precision evaluates within-laboratory variation across relevant routine conditions, such as days, analysts, qualified instruments or replacement columns within the method specification. Reproducibility concerns precision between laboratories.
Common summaries include standard deviation and relative standard deviation (RSD), also called coefficient of variation (CV). RSD is 100 × standard deviation / mean and is useful when the mean is meaningfully above zero. For HPLC area-percent, report the mean, standard deviation in percentage points and, where useful, RSD. For low-level impurities, interpretation must account for reporting and quantitation limits. For signed mass error in daltons or parts per million, report mean error and its standard deviation or a justified interval; RSD becomes unstable or misleading when the mean error is near zero. A categorical identity decision requires suitable agreement or classification measures rather than a CV of pass/fail outcomes.
Published assay evaluations provide examples of separating within-run and between-run performance, although those labels must be interpreted against the actual study design rather than treated as automatic equivalents of repeatability and intermediate precision (PMID:30978327; PMID:34107635). Measurement-uncertainty work also illustrates why precision and calibration contributions should be distinguished (PMID:19269428).
Every precision statement should identify the attribute, preparation count, injection count, conditions and statistical method. A number without that scope cannot establish which sources of variation were assessed.
Which HPLC and LC–MS parameters should a precision protocol record?
A precision protocol should retain the reportable result and the diagnostics needed to interpret it. For reversed-phase HPLC, relevant fields may include retention time, peak symmetry or tailing, peak width or plate number, resolution of a specified critical pair, main-peak area, area-percent, excluded regions and integration events. Define each metric and its purpose before collecting data. Retention-time consistency can support an identification procedure, but does not independently establish chemical identity.
System suitability assesses whether the analytical system meets predefined criteria for its intended use. Depending on the procedure, those criteria may include reference-solution peak-area RSD, retention-time consistency, tailing and resolution. Limits are method-specific. Record whether reference injections come from one stock or independently prepared solutions, because those designs include different variance sources. A failed system-suitability result requires investigation under the laboratory procedure; sample replication alone does not resolve it.
For LC–MS, define whether the precision endpoint is extracted-ion area, calibrated concentration, measured mass, mass error or a relative-response estimate. Isotope-pattern or identification scores require statistics appropriate to the score and decision rule. Relative ion response is not automatically a molar impurity percentage: ionisation efficiency, adduct formation and matrix effects can differ between species. Published high-resolution mass-spectrometric impurity methods are relevant examples of attribute-specific validation (PMID:39778260). Subunit mass analysis provides a separate example of monitoring oxidative variants rather than relying only on ultraviolet area-percent (PMID:28106519).
Document acquisition settings, including wavelength and bandwidth where applicable, sampling rate, mass calibration, extraction windows, deconvolution settings and integration rules. Uncontrolled changes in these settings can make results non-comparable even when the reported RSD appears small.
How should a laboratory design a repeatability experiment for peptide purity methods?
Begin with the intended reportable result and define what constitutes an independent replicate. Repeated injections of one solution estimate injection-level variability. Independent weighings or reconstitutions from a suitably homogeneous lot assess the preparation steps included in the routine procedure. Multiple injections from each preparation can help separate these components, but must not be counted as independent preparations.
Choose concentration levels and replicate numbers according to the method's intended range, reporting threshold and required confidence. Where a named validation framework applies, use its relevant design requirements. A nominal-concentration study may suit one reportable attribute, while a related-substance method may need precision assessment across low and higher impurity levels. Predefine acceptance criteria rather than selecting them after viewing the results.
Randomise or block the run order where practical, while controlling solution age and stability. Bracketing reference injections can help detect drift, and blanks can assess carry-over. Retain preparation order, injection order and elapsed time so a trend is not mistaken for random scatter.
Peptide-specific sources of variation can include incomplete dissolution, adsorption to containers or tips, filtration losses and chemical change during the sequence. Their importance is method-dependent and should be investigated rather than assumed. A progressive fall in parent response or rise in an oxidative variant warrants a stability or handling investigation. Stressed samples may challenge specificity and integration, but do not replace representative samples for routine precision assessment.
Specify the reporting unit, averaging rule, integration method and handling of deviations or exclusions in advance. Absolute-area variability may reflect preparation concentration, injection volume and detector response. Area normalisation can cancel some common scaling effects, but can also conceal losses or co-eluting material and does not establish mass-fraction purity. Do not pool ultraviolet and mass-spectrometric RSDs as though they were measurements of the same attribute.
Which operational factors belong in an intermediate-precision study?
Intermediate precision should represent relevant within-laboratory operating conditions. Candidate factors include day, analyst, independently prepared mobile phase, qualified instrument and replacement columns within the specified chemistry and dimensions. A planned study is usually easier to interpret than a retrospective pool, although suitable historical data can contribute when their structure, traceability and representativeness are justified.
Avoid changing every factor together if the objective is to estimate individual variance components. For example, assigning one analyst exclusively to one day confounds analyst and day effects. Use a feasible crossed or nested design that matches the available resources, with enough replication to support the intended analysis. A small study may estimate combined intermediate precision without reliably ranking individual contributors. State that limitation rather than assigning unsupported causes.
Chromatographic contributors can include mobile-phase composition, pH, column variability, dwell volume and detector response settings. Sample-preparation contributors can include balances, pipettes, mixing and containers. Routine variation within the written method belongs in intermediate-precision planning. Deliberate small changes to method parameters are generally considered through robustness work; findings from that work can inform which conditions need control.
Inspect results by day, analyst, preparation and run order before calculating a pooled summary. Investigate unequal variances, drift and atypical results. Do not discard a poor day merely because it increases the RSD. Any exclusion must follow predefined procedures or a documented, scientifically justified investigation; retain the original data and assess whether the event reveals a routine operational risk.
Published assay evaluations illustrate reporting separate within-run and between-run layers (PMID:30978327; PMID:26555316). Their numerical CV limits should not be transferred to HPLC area-percent or LC–MS measurements without justification. Report the factors studied, independent preparation count, injection count, statistical model and estimated precision. Intermediate precision includes relevant repeatability contributions, so the two estimates should not be interpreted as unrelated quantities.
How should calibration, orthogonal methods and batch files document precision?
Precision of area-normalised HPLC purity is not interchangeable with precision of a calibrated content assay. A calibrated assay may include variability from standard preparation, calibration fitting and response measurement. Reference-value uncertainty can affect measurement uncertainty and trueness without necessarily producing observable scatter in repeated measurements using the same reference stock. Do not label every uncertainty contribution as repeatability.
Calibration strategy should match the measurand and intended range. Peptide-based and protein-based calibration have been compared in isotope-dilution LC–MS/MS work (PMID:39745055); conclusions from that setting should not be assumed to apply unchanged to every peptide method. Characterised in-house references require documented value assignment and uncertainty appropriate to their use. Low-level impurity precision should be assessed at relevant concentrations, using justified handling of results below quantitation or reporting limits.
Orthogonal methods can expose analytical blind spots, but do not automatically improve precision or measure the same quantity. Stable ultraviolet area-percent alongside a changing mass-spectrometric variant signal warrants investigation of selectivity, response and sample stability rather than declaring either method more precise. Subunit mass analysis and high-resolution impurity methods provide relevant analytical examples (PMID:28106519; PMID:39778260). Uncertainty estimation in dual-detector size-exclusion chromatography and mass-spectrometric protein fingerprinting offer further methodological context, not universal peptide acceptance limits (PMID:19269428; PMID:31759609).
A research certificate of analysis is not a complete validation or verification file. It should identify the lot, method and reportable results under the laboratory's quality system. If it reports precision, distinguish independent preparations from repeated injections and state the relevant n. System-suitability results may appear on the certificate or in linked records as required. Do not imply that a full intermediate-precision study was repeated for every lot. Assay evaluation papers are examples of reporting practice, not substitutes for method-specific criteria (PMID:34107635; PMID:30978327).
The supporting method file should retain the protocol, acceptance criteria, factor design, raw data, chromatograms, integration history, statistical analysis and deviation records. Useful traceability fields include method version, column lot and serial number, instrument and analyst identifiers, preparation and sequence identifiers, raw-data links, weigh-in, diluent, acquisition settings and reporting threshold. Use consistent lot identifiers or a controlled cross-reference between the physical container, certificate and analytical sample records.
Changes to column chemistry, detector configuration, software or integration defaults require documented impact assessment. Depending on their effect, targeted verification or broader revalidation may be needed; a change does not automatically invalidate every historical precision result. System-suitability trending can support ongoing monitoring but does not replace assessment of sample-preparation performance. These records support analytical characterisation of laboratory research reagents only, not suitability for human or veterinary administration.
Source materials that match this documentation standard
The sections above describe how serious laboratories evaluate identity, purity, and batch records. When you are ready to source research materials against that same standard, ClaraScience supplies from Australian warehouses with Express tracked dispatch and batch documentation on every order.
Retail catalogue orders ship with lot documentation. Qualified buyers can request wholesale portal access for bulk restocks and tier pricing.
Frequently asked questions
Is repeatability the same as intermediate precision for a peptide HPLC method?
No. Repeatability concerns a short interval under the same operating conditions. Intermediate precision includes relevant within-laboratory variation across conditions such as days, analysts or instruments. Repeated injections of one solution assess injection-level repeatability; independent preparations are needed to include sample-preparation variability. State the scope and replicate structure for each estimate.
Should every research certificate of analysis include a sample RSD?
Not necessarily. Certificate requirements depend on the method and quality system. If an RSD is reported, identify the measured attribute, the number of independent preparations and the number of injections. Full precision-study records belong in the supporting method file. A single chromatogram is not a precision study, and system-suitability RSD is not automatically sample-method RSD.
Why is related-substance RSD often larger than main-peak area-percent RSD?
At lower responses, baseline noise, integration and ionisation variability can represent a larger fraction of the result. Area normalisation can also reduce some common scaling variation in the main-peak percentage. Set criteria by attribute, concentration and technique. Near zero or below the quantitation limit, RSD may be unstable or uninformative, so use a justified statistical and reporting approach.
Can immunoassay CV limits be copied onto a peptide HPLC purity method?
No. Immunoassay studies can illustrate how to report different precision levels, but their numerical limits are platform- and purpose-specific. HPLC area-percent, LC–MS response and ligand-binding concentration are different measurands or measurement processes. Acceptance criteria need justification for the actual method, range and intended research use.
What is the difference between SST precision and sample repeatability?
System-suitability precision evaluates specified system performance, often through repeated reference-solution injections. Sample repeatability evaluates results for the test material under repeatability conditions and should include preparation steps when that is the intended claim. Passing SST does not by itself establish sample-preparation precision, specificity or the reliability of every reported attribute.
Which documentation fields link precision data to a research vial?
Link the lot and analytical sample identifiers to the method version, preparation records, instrument and analyst identifiers, column records, sequence, raw-data files, integration history and reported results. Record relevant weigh-in, diluent, acquisition settings and reporting thresholds. Use consistent identifiers or a controlled cross-reference between the vial, certificate and analytical records. Shipping location alone does not establish analytical traceability.
References
- PMID:30978327 — C-peptide and insulin assays with the Mindray CL-2000i: Precision and comparability with different methods — Clin Chim Acta — 2019
- PMID:34107635 — Analytical Evaluation of Lumipulse Anti-Mullerian Hormone Assay — Clin Lab — 2021
- PMID:19269428 — Estimation of uncertainty in size-exclusion chromatography with a double detection system (light-scattering and refractive index) — Talanta — 2009
- PMID:39778260 — Validation of a liquid chromatography-high-resolution mass spectrometry method to quantify peptide-related impurities in teriparatide — J Pharm Biomed Anal — 2025
- PMID:28106519 — Subunit mass analysis for monitoring antibody oxidation — MAbs — 2017
- PMID:26555316 — Technical performance of a novel, fully automated electrochemiluminescence immunoassay for the quantitation of β-amyloid (1-42) in human cerebrospinal fluid — Alzheimers Dement — 2016
- PMID:39745055 — Comparative evaluation of peptide vs. protein-based calibration for quantification of cardiac troponin I using ID-LC-MS/MS — Clin Chem Lab Med — 2025
- PMID:31759609 — Protein fingerprinting and quantification of β-casein variants by ultra-performance liquid chromatography-high-resolution mass spectrometry — J Dairy Sci — 2020
Research use only
This article is provided for laboratory research and educational purposes only. Products referenced are not for human or veterinary use. ClaraScience makes no therapeutic, medical, or efficacy claims, and nothing here constitutes medical advice.
