1. Preserve and validate the CSV source
Upload the authorized bounded UTF-8 CSV and retain its unchanged bytes, filename, media type, server-computed SHA-256 identity, import time, and reader profile. Reject an unsupported encoding, size, shape, or response value explicitly.
Inspect the detected or selected measurement region before mapping. Reader exports may contain metadata, headers, multiple rectangular blocks, or calculated values; an absent or ambiguous plate region must stop the workflow rather than lock the wrong numbers into an analysis.
2. Assign scientific meaning to each well
Use a visual grid to assign sample, role, replicate, dilution level, and inclusion state. Bulk mapping speeds repeated layouts; selected-well editing handles exceptions. Both should produce the same structured records.
The four-well slice below shows the minimum bridge from a rectangular reader export to analysis-ready records. Continue the same mapping through all 96 wells, including blanks and unused positions, so no measurement disappears because its label was blank.
| Well | Sample | Role | Replicate | Dilution level | Raw response |
|---|---|---|---|---|---|
| A1 | Reference | Reference | 1 | 1 | 1,742 |
| A2 | Reference | Reference | 2 | 1 | 1,716 |
| B1 | Reference | Reference | 1 | 2 | 1,406 |
| B2 | Reference | Reference | 2 | 2 | 1,392 |
3. Derive concentrations in a reviewable table
For each sample, record initial concentration, dilution factor, number of levels, and unit. Generate concentration from those inputs while allowing a method-defined explicit series when dilutions are not uniform.
The long-form table should bring the calculated concentration beside the original well response. This is the bridge between the physical plate and the statistical model.
concentration(level n) = initial concentration ÷ dilution factor^(n − 1)4. Review controls, replicates, and plate position
Evaluate required controls, missing wells, response range, and replicate mean, sample SD, and CV while preserving individual values. Inspect position-aware views for row, column, edge, timing, or dispense patterns.
When the procedure permits an exclusion, require a reason and retain the original well and value. The locked analysis must identify exactly which observations are included; an excluded well never disappears from review or report evidence.
5. Run the approved analysis and suitability checks
Apply the approved method’s exact model contract, transformations, exclusions, diagnostics, and suitability criteria. Display observed and fitted data, parameters, potency and uncertainty, warnings, and each criterion’s observed value and outcome.
A failed, not-evaluable, or unconverged result remains visible. The analysis produces the controlled outcome; the laboratory procedure determines any investigation, repeat, or disposition.
- Lock the reviewed source region, well map, dilution series, and inclusions.
- Fit only the model and settings named by the method.
- Inspect plotted observations, residuals, parameter state, and warnings before reading the reported result.
- Evaluate every configured criterion from its retained observations.
- Generate the report from the same locked analysis rather than copying a displayed value into a second document.
6. Generate the retained source-to-result report
Connect the final result to the source fingerprint, selected region, well mapping, dilution table, individual responses, exclusions and reasons, method and configuration versions, model output, suitability, analysis release, and review context.
Generate the report from the retained analysis values. A downstream copy or transcription should preserve the report identity rather than becoming an unexplained second result.
Limits and when a controlled workflow helps
Reader formats are not interchangeable. Confirm the exact export layout and measurement region against representative valid, boundary, and invalid files, and control any conversion before analysis.
Provenarium imports supported bounded CSV, TSV, and TXT profiles and curated static XLSX (Excel) workbooks with cached formula results. It does not execute formulas or macros, read proprietary instrument files, or acquire data directly from instruments. For a matching profile, it can keep plate mapping, approved analysis settings, suitability, exclusions, and the resulting report in one controlled routine workflow.
Frequently asked questions
What if the reader export is an Excel workbook?
Provenarium can accept a curated static XLSX workbook when required formula results are already cached. It does not calculate workbook formulas or run macros. Retain the authoritative workbook and reconcile the selected measurement region, wells, and values before analysis.
Should replicate wells be averaged before curve fitting?
That is a method decision. Preserve every well either way, document the summarization rule, and ensure uncertainty calculations match the unit of analysis.
How should excluded wells appear in the report?
Identify the well, original value, reason, and effect on inclusion. Never delete or overwrite the source observation.