Separate source preservation from analytical structure
Reader exports are often optimized for instruments or human viewing rather than statistical analysis. Preserve the original file and selected sheet, then create a long-form analytical table with one record per well or observation.
At minimum, retain run and plate identifiers, well, raw response, sample, role, replicate, dilution level, calculated concentration, inclusion state, and exclusion reason. Derived values should identify their source and transformation.
| Field | Purpose |
|---|---|
| well | Trace the value back to physical position |
| sample and role | Identify reference, test, standard, control, or unknown |
| replicate and dilution | Reconstruct the experimental design |
| raw response | Preserve the instrument measurement |
| concentration and unit | Make dilution calculations explicit |
| included and reason | Keep exclusions reviewable |
Treat mapping as data, not decoration
A visual plate makes errors visible, but the saved mapping is the analytical input. Bulk assignment, named templates, validation checks, and a selected-well editor should all write the same structured mapping.
Validate duplicate wells, missing roles, invalid dilution indices, unexpected blanks, and incompatible sheet dimensions before fitting. Display unmapped measurements rather than silently dropping them.
Look for spatial and process effects
Rows, columns, edges, timing, dispensing, and reader behavior can introduce systematic patterns. A plate heatmap, control-location review, and replicate summaries can reveal issues that disappear in a curve-only view.
A visible pattern is a diagnostic signal, not an automatic exclusion rule. Any exclusion should identify the observation, reason, actor or source, and downstream impact.
Frequently asked questions
Should the original spreadsheet be modified before import?
Prefer a documented reader profile or mapping step that retains the source file unchanged. Manual cleanup can be practical, but it should not erase how the analytical table was derived.
Can one importer support every plate reader?
Instrument and laboratory exports vary. Use named, tested import profiles for the formats your laboratory actually receives, and verify each profile against representative exports. Universal compatibility should not be assumed.