Start with the original measurement and context
Retain the original reader output or a justified true copy with its source data file fingerprint, import time, instrument/run identity, selected sheet or channel, and reader profile. A pasted matrix without its source context may be insufficient to explain how values entered the analysis.
Make user and automated actions attributable. Service accounts and background calculations need identities and versions just as human edits need users and timestamps.
Keep mapping and concentration calculations reviewable
The plate map assigns scientific meaning to positions. Preserve sample names, roles, replicates, dilutions, controls, reference designation, initial concentrations, factors, and units in the versioned analysis configuration.
Derived long-form data should point back to the source well. If a mapping changes, preserve the prior state and reason where the workflow requires change history.
Never make exclusions disappear
An excluded well should retain its raw and derived values, position, reason, user or source, time, and relationship to the final analysis. Reports should disclose exclusions and their rationale at a useful level.
Conditional formatting can draw attention to failures, but it is not a data-integrity control. Store the observed value, rule, outcome, and display style separately so changing the report format does not change the scientific meaning.
Translate each ALCOA+ principle into a concrete record
The principles become useful when each one points to something a reviewer can inspect. The examples below are starting questions for a plate-reader workflow, not a substitute for defining the complete record and controls for your intended use.
| Principle | Plate-reader analysis example |
|---|---|
| Attributable | User, service, instrument, and software identity for each recorded action |
| Legible | Readable source values, settings, outcomes, reasons, units, and timestamps |
| Contemporaneous | Import, mapping, exclusion, analysis, and review recorded when performed |
| Original | Retained reader output or justified true copy linked to derived values |
| Accurate | Tested import, calculation, and transfer with visible corrections |
| Complete | Included and excluded data, failed runs, settings, history, and reports |
| Consistent | Ordered timestamps, stable identifiers, units, and version relationships |
| Enduring | Records kept in durable, protected storage for the required period |
| Available | Complete records can be retrieved, rendered, and inspected when needed |
Define the complete analysis record
The complete record may include source files, parsed data, plate map, dilution table, analysis settings, software version, parameters, diagnostics, system-suitability outcomes, exclusions, report settings, generated report, fingerprints for the source data file and analysis configuration, analyst and date information, and review decisions.
Retention, availability, backup, restoration, permissions, audit trail, and downstream reconciliation need operating controls beyond generating a PDF. Evaluate those controls for intended use.
Frequently asked questions
Is a SHA-256 fingerprint proof that data are correct?
No. A fingerprint can help detect whether bytes changed and link related records, but it does not prove the source was accurate, complete, properly attributed, or reviewed.
Is the PDF the original record?
That depends on the defined record and workflow. A PDF may be only one part of the record; source files, structured data, analysis settings, and history may also need to be retained.