Importing and validating fluorescence readings from plate readers

What to check before your raw OD or RFU values ever touch an analysis pipeline

What the plate reader gives you

A plate reader exports a matrix of numbers, one per well. Depending on your instrument and assay, those numbers are:

FormatUnitCommon assay
AbsorbanceOD (optical density, unitless)ELISA, protein quantification
Fluorescence intensityRFU (relative fluorescence units)Cell viability, reporter assays
LuminescenceRLU (relative light units)ATP assays, luciferase reporters
Time-resolved fluorescenceCounts or ratioHTRF, TR-FRET

All of these are relative measurements. They only mean something in the context of controls on the same plate.

Validating the raw file

Before importing, open the exported file and check for these common problems:

Merged or missing headers

Plate reader software often exports files with merged cells, multi-row headers, or instrument metadata above the data block. Platelet handles most common formats automatically, but verifying manually the first time saves debugging later.

Saturated wells

A well reading at or very near the instrument's maximum value is saturated: the detector cannot distinguish between a strong signal and an overwhelming one. Typical saturation thresholds:

FormatSaturation indicator
AbsorbanceOD > 3.5 (most readers)
FluorescenceInstrument max, often 65535 counts
LuminescenceOverflow flag in exported file

Saturated wells should be excluded before any downstream analysis.

A saturated positive control makes your Z′-factor calculation meaningless. If your positive control is saturated, dilute it and re-run rather than proceeding with normalisation.

Checking the signal window

The signal window is the ratio of your positive to negative control means. A window below 3× is generally too small to detect weak hits reliably.

Check your signal window plate-by-plate, not as a global average across a run. A single plate with a collapsed window should be flagged and re-run, not silently averaged away.

Plate-to-plate consistency

If you are running multiple plates in a single experiment, compare the negative control means across plates before merging the data. A drift of more than 20% between plates indicates a systematic problem, such as reagent degradation, incubation time differences, or reader lamp variation.

Importing into Platelet

Platelet accepts Excel and CSV exports from all major plate reader software. Upload the file in the Readings tab, map the data block to the correct wells, and the app will flag saturated values and compute the signal window automatically before you proceed to analysis.

Ready to run this analysis on your own data?