ELISA Troubleshooting: A Low Standard Curve and Samples Reading Off-Scale

The problem

A researcher measuring Periaxin (PRX) in human serum reported that the assay could not quantify any of the samples: every well read above the top standard. Two things stood out on the first run. First, the sample optical densities were very high — roughly 2.1–3.0 at 450 nm — above the highest calibrator, so no concentration could be interpolated. Second — and just as important — the standard curve itself was clearly deficient: the top standard reached only OD ≈ 1.5 against an expected ≈ 2.0, so the whole curve was compressed into a narrow, shallow signal window, and the two lowest standards read at or below the blank, leaving the bottom of the curve with no working resolution at all. A curve in this state cannot support a reliable interpolation, whatever the samples happen to read — so it has to be fixed before anything else. The initial suspicion was a faulty kit or standard.

Snapshot of the data

Two diagnostics tell the story: the shape of the standard curve, and where the samples sit relative to it. Look at the curve first, because it is the reference everything else is measured against. On the first run it was visibly deficient — the top standard topped out low, compressing the dynamic range, and the two lowest standards did not separate from the blank, so the entire bottom of the curve was flat and unusable. Only once that is understood do the sky-high sample readings make sense.

Standard curve — first run (duplicate ODs at 450 nm)

Standard (ng/mL)	OD rep 1	OD rep 2	Separates from blank?
10 (top)	1.394	1.696	Low — expected ≈ 2.0
5	0.923	1.028	Yes
2.5	0.505	0.572	Yes
1.25	0.313	0.330	Yes
0.625	0.245	0.240	Barely
0.312	0.231	0.203	No — at blank
0.156 (lowest)	0.204	0.189	No — below blank
Blank	0.216	0.186	—

Sample readings vs. the top standard

Attempt	Dilution	Typical sample OD (450 nm)	Interpretation
Run 1	Undiluted (neat)	≈ 2.1–3.0	Above top standard — over range
Run 2	1:6	≈ 2.8–3.1	Still above top standard

Standard curve — repeat run with a freshly reconstituted standard

Standard (ng/mL)	OD rep 1	OD rep 2	Separates from blank?
20 (top)	2.162	2.203	Reaches ≈ 2.2
10	1.774	1.732	Yes
5	1.166	1.007	Yes
2.5	0.606	0.546	Yes
0.312	0.166	0.152	Yes
0.156 (lowest)	0.134	0.136	Yes — above blank
Blank	0.125	0.113	—

Customer, institution and sample identities are anonymized; values are illustrative of the reported pattern. The repeat curve returned an excellent fit (R² ≈ 0.999).

Why it happens

A sample reading above the top standard means only one thing directly: the signal has saturated beyond the calibrated range, so the true concentration cannot be read. In serum, this is common when the target is abundant and when matrix components — proteins, lipids and other interferents in neat or lightly diluted serum — add to the signal. A matrix effect occurs when substances in the sample interfere with the assay, producing artificially high (or occasionally low) readings. Diluting the sample both lowers the analyte into range and dilutes the interferents, which is why a too-modest dilution such as 1:6 can leave readings pinned at the top of the scale.

Our analysis

Two separate issues were tangled together, and separating them was the key — starting with the standard curve, because until the curve is sound, no sample value can be trusted. On the first run the curve was genuinely deficient: the top standard fell well short of the expected OD ≈ 2.0, compressing the whole curve into a shallow signal window, while the two lowest standards collapsed onto the blank, leaving the low end with no working range at all. A curve in this condition cannot anchor a reliable back-calculation — even a perfectly diluted sample would be read against an untrustworthy scale. That signature almost always points to a standard reconstitution or handling problem rather than the assay chemistry itself. And that is exactly what we saw: when the researcher rebuilt the curve with a freshly reconstituted standard, it came back textbook — the top point reached OD ≈ 2.2, every standard sat cleanly above the blank, and the fit returned R² ≈ 0.999 — confirming the kit itself was performing correctly and that the first curve had simply been prepared poorly.

That left the real problem in plain view. Even at 1:6, the samples read around OD 3.0 — higher than the top standard and, in fact, higher than the neat readings from the first run. A kit validated on serum at 1:1000 was being asked to read serum at 1:6, roughly a hundred-fold too concentrated. The samples were simply far over range, and the dilution had to move much closer to the validated factor.

Root cause

What we recommended

The path back to quantifiable data was to confirm the curve, then bracket the correct dilution:

• Reconstitute the standard carefully — briefly centrifuge the vial, add the specified volume of Sample Diluent, and let it stand ~15 minutes with gentle mixing before preparing the serial dilutions.
• Confirm the curve before trusting samples: the top standard should reach OD ≈ 2.0 and every standard, including the lowest, should sit clearly above the blank.
• Bracket the dilution: run a small subset at 1:10, 1:100 and 1:1000 to find where the sample OD lands mid-curve. Here the answer was near 1:100–1:1000, not 1:6.
• Keep every reported reading inside the linear region of the curve; anything above the top standard must be diluted further and re-run.
• Conserve diluent with serial dilutions — prepare the first dilution in Sample Diluent (which buffers matrix effects), then take the higher dilutions onward in PBS, and plan volumes ahead (a full plate at 1:100 needs roughly 16 mL for samples plus ~7 mL for the standards).

Key takeaway

Frequently asked questions

My standard curve tops out low — what does that mean?

If the highest standard falls well below the expected OD (around 2.0 for this kit), the curve is compressed and the low end may collapse onto the blank. This usually reflects standard reconstitution or handling rather than a defective kit. Re-reconstitute the standard carefully and re-run before trusting any values.

My samples read higher than the top standard — is the kit broken?

Usually not. A reading above the top calibrator means the analyte is over range and cannot be interpolated. Dilute the sample so the signal falls inside the curve, then re-run.

How do I find the right dilution?

Run a small pilot at 1:10, 1:100 and 1:1000 and choose the dilution that places the sample OD in the middle of the curve. Then use that factor for the full plate.

Can I dilute in PBS to save Sample Diluent?

Prepare the first dilution in Sample Diluent, which contains components that buffer matrix effects. Subsequent, higher dilutions can be made in PBS — as the sample is diluted further, the matrix effect naturally decreases too.

Written by the Assay Genie scientific team — reviewed for accuracy by our PhD scientists. Customer, institution and sample identities have been anonymized; data values are illustrative of the reported pattern and provided for educational purposes.