What do data integrity (ALCOA+) principles require in laboratory peptide documentation?
Attributable means the person or system that created or changed a record can be identified: the analyst who prepared the sample, the reviewer who accepted the integration, and the instrument account that acquired the chromatogram or spectrum. A shared login with no user mapping fails attribution even when the peak table looks tidy. Legible means a qualified reviewer can read peak codes, units, wavelengths and lot identifiers without private shorthand, using the laboratory's catalogue of abbreviations. Contemporaneous means the entry is made at the time of the activity. A purity figure typed into a spreadsheet days later, with no link to the acquisition time stamp, is a summary, not the original observation. Original means the first-capture record is retained — the chromatography data-system result, the mass raw file, or a controlled true copy that preserves content and meaning. A photograph of a monitor is not an original file. Accurate means the value reflects the observation and the figure on the certificate matches the stored peak table.
The plus attributes commonly add complete, consistent, enduring and available. Complete means blanks, system-suitability runs, sample runs and the processing method are present, not only a selected trace. Consistent means lot identifiers, units and retention-time windows do not drift between vial, sequence and certificate. Enduring means the record stays readable for the laboratory's stated retention period, including after a software change. Available means an authorised reviewer can retrieve the record and its audit trail in the time the quality system specifies. A 2026 review examines data integrity in pharmaceutical quality assurance through the ALCOA principles (DOI:10.33545/26647222.2026.v8.i1b.300). These attributes do not describe any effect of a peptide in a person.
How are HPLC and mass-spectrometry records made attributable and contemporaneous?
A chromatography sequence is a record, not a queue. Each row should carry a sample name that matches the container label, a vial position, a method identifier, an acquisition date and time from a synchronised clock, and the instrument identity. The processing method — integration events, detection channel, and any manual baseline change — belongs with that row. If an analyst reprocesses a peak, the audit trail should show who changed the event, when, and why. A reason code such as adjusted, with no description, does not let a second reviewer judge whether the change was scientific or cosmetic. If the balance clock, the data system and the worksheet disagree by hours, the file cannot show that weighing, dilution and acquisition occurred in the order the worksheet describes.
Mass records need the same discipline. A monoisotopic mass copied onto a certificate without the spectrum, the charge-state assignment and the calibration or tune-file identifier is a transcription. The original is the raw or controlled processed spectrum for that lot, stored so the ion list can be reopened. Where tandem fragmentation supports a sequence assignment, the record should identify the precursor, the fragment set interpreted, and the person who accepted the assignment. Pasting a theoretical mass beside an unrelated spectrum is an accuracy failure even if the numbers are close.
A signature is meaningful only if the signed object is defined: the result set, the peak table, or the certificate. Signing a PDF while the project remains editable, with no lock or hash recorded at sign-off, leaves a gap between approval and the original. Where one person holds more than one role, the procedure should say so, and the audit trail should still show each action. These fields reconstruct what was measured. They are not instructions for any use of the material outside the analytical method.
Is traceability the link across ALCOA, ALCOA+ and ALCOA++ in a peptide batch file?
A 2022 discussion asks whether traceability is the glue for ALCOA, ALCOA+, or ALCOA++ (DOI:10.56530/spectroscopy.up8185n1). Each attribute can look satisfied on a single page and still fail when the pages do not join. Traceability is the chain from the container in hand to the original acquisition and back, without relying on memory.
A workable chain has explicit joins. The lot identifier on the vial should match the lot identifier on the sample-preparation worksheet. The worksheet should name the balance, the mass weighed, the volumetric ware, the diluent and the preparation time, and it should point to the sequence row. The sequence row should point to a data-file name or result identifier in the chromatography or mass-spectrometry system. That identifier should appear on the peak table and the certificate header. The reviewer sign-off should name the result set reviewed, not an undated summary. If a join is only verbal, the record is not traceable, however legible each page is.
Changes belong on the same chain. A corrected mass, a reintegrated peak or a replaced certificate should point to the superseded version and the reason. A controlled true copy is acceptable when the procedure defines how scale, peak markers and metadata are preserved. An image that omits the y-axis, the method identifier or the sample name breaks the chain.
An ALCOA++ discussion of pharmaceutical digital twins and continuous manufacturing integrates human factors and simulation vulnerabilities into the integrity picture (DOI:10.1007/s12247-026-10778-6). A simulated trace, a method copied from another lot, or a digital-twin output must be labelled as such. It must not be filed as the original acquisition for the container placed into research stock. Traceability fails when a model product is substituted for a measurement.
What should a complete, consistent and enduring peptide batch file contain?
Completeness fails when a certificate is detached from its data. A reviewable research file holds separate objects: the certificate with specification, result and disposition; the peak table with retention time, area, area percent and peak labels; the chromatogram with axes, sample name and method identifier; system-suitability evidence for that sequence; and, where identity is claimed by mass spectrometry, the spectrum or a controlled extract with mass and charge state. If related substances are reported, the table should state how peaks below the reporting threshold were treated and whether solvent or blank peaks were excluded. Silent exclusion makes an area-percent total look closed when it is not. System-suitability evidence should name the parameters the method actually uses, such as retention-time agreement across bracketing runs and a recorded peak-symmetry or plate-count check, together with the laboratory's own acceptance values. Those values belong in the file; they should not be assumed from a generic template.
Consistency is arithmetic and vocabulary. Reported area percentages, after documented exclusions, should reconcile to the purity statement within the laboratory's rounding rule. An area-normalised figure is not an assay, and an assay against a reference standard is not chromatographic purity. Counterion content, water content and net peptide content are further measurements, each with its own method identifier and unit. Folding them into one unqualified purity word misleads a laboratory comparing lots. A mass unit on the spectrum and a different unit on the certificate, with no conversion shown, is not reviewable.
Raw files should be kept with the software name and version needed to open them. A PDF can be a controlled copy if the procedure states what it preserves; it is a weak sole archive if scale, integration events and audit history are lost. The index key should be the lot identifier a reviewer will search, not only an instrument run number buried in an image. A 2024 paper explores industry challenges and solutions for data-security resilience in AI-driven digital transformation through ALCOA+ principles (DOI:10.5455/aim.2024.32.65-70).
Accuracy also includes the validated interval. A 2023 abstract on data integrity and uncertainty asks how uncertainty is derived from method validation and how a result is interpolated on the validated interval (DOI:10.1111/jvp.13209). Extending a calibration past that interval, and still calling the figure the validated method, overstates the record. Area percent is not a calibrated content result. Where content is reported, the method identifier and range should travel with the number. A bare figure with no range cannot be checked.
How should a laboratory judge automated scores and supplier batch files?
Tools that score compliance are now part of the published discussion. A 2023 paper describes deep transformers for computing and predicting ALCOA+ data-integrity compliance in the pharmaceutical industry (DOI:10.3390/app13137616). A predicted label is still a secondary computation. It is attributable only if the model version, the input files and the person who accepted the output are recorded. It is accurate only for the features the model saw. If the model never opened the raw spectrum, the score cannot stand in for identity confirmation.
Human-factor failures remain the ordinary route to a legible but unreliable file. Examples include a shared password, a worksheet completed after the sequence, a manual integration used to hide a shoulder without a documented reason, and a result transcribed from a screen by memory. Simulation failures look different and have the same effect. A processing method copied from another lot, a theoretical isotope pattern pasted as if acquired, or a preview trace filed in place of the run, all produce a page that is not the original. Reviewers should ask which file was acquired from the vial, not which file looks finished.
Unauthorised change, silent overwrite and loss of the audit trail are integrity failures even if the numbers were right at acquisition. Unique accounts, synchronised clocks, restricted deletion, backup with a restore check, and a defined certificate export should be written into the procedure. An automated score may help prioritise human review. Discarding the original because a model marked the file compliant inverts the principle.
An Australian receiving laboratory can apply the same tests before a lot enters research stock. Match the lot on the label to the certificate header and the chromatogram sample name. Confirm a result identifier, not only a cropped image. Check that method, detection channel and peak-table arithmetic are present and that the stated figure reconciles. Prefer a file that remains readable, indexed by lot, over a link that may expire. Tracked dispatch records and the batch file should carry the same lot identifier. That check is a procurement control for research material. It is not an assessment of any human use.
Frequently asked questions
What does ALCOA stand for in a peptide laboratory record?
Attributable, legible, contemporaneous, original and accurate. In laboratory peptide documentation those words test whether a chromatogram, spectrum or certificate can be tied to a person or system, read by a reviewer, placed in time, traced to first-capture data and shown to match the calculation. The plus attributes commonly add complete, consistent, enduring and available. Together they describe record quality for research lots. They do not describe a biological effect.
What is the difference between ALCOA, ALCOA+ and ALCOA++?
ALCOA is the five-term core. ALCOA+ is the common extension that adds complete, consistent, enduring and available, so a file is judged on what it contains and whether it can still be retrieved. ALCOA++ is used in some manufacturing and digital-system discussions, including work that places human factors and simulation risk beside the original attributes. A spectroscopy discussion asks whether traceability is the glue across all three labels. Laboratories should define the terms in their own procedure rather than assume every author uses the same extra letters.
Is a PDF of a chromatogram an original record?
Not by itself. The original is the first-capture file in the chromatography or mass-spectrometry system, or a controlled true copy that preserves content and meaning, including scale, sample name, method identifier and processing history. A PDF may be an acceptable copy if the procedure says what it keeps and how it is linked to the result identifier. A cropped image with no axes and no lot identifier fails legibility and traceability even when the peak shape is clear.
Can software predict ALCOA+ compliance for a peptide file?
Published work has explored deep-learning models that compute or predict ALCOA+ compliance from pharmaceutical data. A score of that kind is a secondary record. It should name the model version, the inputs and the person who accepted it. It cannot replace the raw chromatogram or spectrum, and it should not be the only file retained. Use it, if at all, to prioritise human review of audit trails and metadata, not as a release decision for a research lot.
Does a strong data-integrity file imply the peptide is suitable for people?
No. Data integrity (ALCOA+) principles in laboratory peptide documentation govern how analytical records are created, changed and kept. A complete, attributable file can support identity, purity and content checks for research procurement. It says nothing about effect, suitability for a person, or any use outside the laboratory method. Research material should be documented and handled as research material. A certificate that adds therapeutic language is a documentation problem, not extra evidence.
What should an Australian laboratory match before accepting a supplier lot?
Match the lot identifier on the container, the certificate header, the chromatogram sample name and any tracked dispatch record. Look for a result identifier that points back to the data system, a method identifier, a detection channel and a peak table whose arithmetic agrees with the stated figure. The file should remain readable and be indexed by lot. Absence of those joins is a traceability gap. The check qualifies the record for research stock. It does not qualify the material for any other use.
References
- DOI:10.33545/26647222.2026.v8.i1b.300 — Data integrity in pharmaceutical quality assurance: A review based on ALCOA principles — International Journal of Pharmacy and Pharmaceutical Science — 2026
- DOI:10.1007/s12247-026-10778-6 — Data Integrity Failures in Pharmaceutical Digital Twins and Continuous Manufacturing: An Alcoa + + Framework Integrating Human Factors and Simulation Vulnerabilities — Journal of Pharmaceutical Innovation — 2026
- DOI:10.5455/aim.2024.32.65-70 — Enhancing Data Security Resilience in AI-Driven Digital Transformation: Exploring Industry Challenges and Solutions Through ALCOA+ Principles — Acta Informatica Medica — 2024
- DOI:10.3390/app13137616 — Deep Transformers for Computing and Predicting ALCOA+Data Integrity Compliance in the Pharmaceutical Industry — Applied Sciences — 2023
- DOI:10.56530/spectroscopy.up8185n1 — Is Traceability the Glue for ALCOA, ALCOA+, or ALCOA++? — Spectroscopy — 2022
- DOI:10.1111/jvp.13209 — O43 | Integrity of data (<scp>ALCOA</scp>) and uncertainty: How derivate the uncertainty from analytical validation method and how interpolate it on the validated interval? — Journal of Veterinary Pharmacology and Therapeutics — 2023
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.
