A Troubleshooting Guide to Blanks, LOB/LOD, and Background-Driven Floors
You ran a clean plate. The pipetting was careful, the washes consistent, the top of the curve textbook. Then you look at the low end and your three lowest standards are sitting almost on top of the blank. Your ELISA standard curve, which should keep descending, flattens out instead, and your low samples come back “below range.”
It’s one of the most common frustrations in immunoassay work, and it’s easy to read as a mistake. It usually isn’t. That flat bottom is the assay telling you where its floor is, and learning to read it is the difference between chasing a problem you can’t fix and fixing the one you can.
What the Shape of Your ELISA Standard Curve Tells You
A well-behaved ELISA standard curve is sigmoidal: low signal at zero analyte, a roughly linear middle, and a high plateau where signal saturates. Fit it with a four-parameter logistic model and the parameters describe that shape, a lower asymptote (A), a slope (B), a midpoint (C), and an upper asymptote (D).
The one that trips people up is the lower asymptote, A. It’s tempting to read it as “zero analyte.” It isn’t. A is your background, the signal the plate gives off when there is essentially nothing to detect, from the well surface, the sample matrix, and the reagents, not your target. So the flat bottom isn’t where the assay ran out of analyte. It’s where the signal has sunk into the background and the two can no longer be told apart, a floor made of noise.
Your Detection Limit Lives in the Blank
Here is the reframe that matters: your sensitivity is defined by the blank, not by how bright your top standard is.
The definitions all point the same way. The Limit of Blank is the highest signal you would expect from a sample with no analyte, by the CLSI EP17 convention the mean of your blank replicates plus 1.645 times their standard deviation. The Limit of Detection builds on it: the Limit of Blank plus 1.645 times the standard deviation of a low-concentration sample. And the Lower Limit of Quantitation, where your trustworthy range actually starts, is the lowest concentration you can measure with acceptable precision, commonly where the CV stays at or below about 20 percent.
Every one of those depends on the blank and its scatter, not the analyte. If your blank reads high or bounces around well to well, your floor rises with it and your lowest standards get swallowed. It’s why two labs running the same kit report different low-end performance: one has a quieter, more consistent blank. So if your low standards are piling up on the blank, the fix is rarely “add more analyte.” It’s “find what’s making my blank loud, and quiet it.”
What Quietly Raises Your Floor
A few usual suspects push the background up. Non-specific binding is the big one; in a plate assay that means hydrophobic adsorption of proteins onto the well surface, heterophilic-antibody interference, and detection antibody catching off-target sites. Matrix effects come next, and they hit the samples that matter most: plasma and serum raise background far more than a clean buffer standard, so a curve that looks beautiful in buffer can collapse on real samples. Then the process levers, incomplete blocking, inconsistent washing, too much detection antibody or enzyme conjugate, aging reagents, each nudging the baseline up.
None are exotic; together they’re a checklist. Before deciding your assay can’t reach lower, walk it, because most “we can’t detect it” problems are really “the blank is too loud” problems.
Why More Signal Isn’t More Sensitivity
Here is the part that surprises people: amplifying your signal does not automatically lower your detection limit.
There are good chemistries for boosting signal, like poly-HRP or tyramide amplification, and they can add ten to a hundred times more signal. But sensitivity isn’t how tall the signal is; it’s how far it sits above the noise. If amplification raises your specific signal and your background in the same proportion, the gap that sets your detection limit hasn’t moved. It can even narrow, because non-selective amplification builds up at non-specific sites too. A brighter plate, the same floor. What lowers a detection limit isn’t raw signal, it’s a better ratio of signal to background.
How to Actually Lower the Floor
So the order of operations is clear. First, quiet the background: optimize blocking, wash more consistently, right-size your detection antibody, handle matrix deliberately. Each tightens the noise band your floor sits on, and a tighter floor is a lower one.
Then, once the background is genuinely quiet, choose amplification that lifts specific signal without dragging background up with it, selectivity, not volume. Nucleic-acid-based signal amplification, the principle behind Cavidi’s Exazym® and its BOLD (Binding Oligo Ladder Detection) technology, is designed to raise specific signal without proportionally raising background, so the gain shows up as a better signal-to-background ratio rather than a taller peak. It is one option among several, and it only helps once the fundamentals are sound. And if your blank is already clean and you still need lower, the honest answer may be a different format or platform, or simply more sample, not more tweaking.
Read the Floor Before You Fight It
Next time your ELISA standard curve flattens at the bottom, read it instead of fighting it. Map your blank and its standard deviation, because that sets your detection limit. Work the background checklist, because a loud blank is almost always the culprit. Then reach for amplification with the right question: not “how do I make more signal,” but “how do I widen the gap between signal and noise.” Do that, and the flat bottom stops being a wall. It becomes a gauge you know how to move.
Want to learn more about Exazym®?
- Explore Exazym® reagent kits
- See how BOLD signal amplification works
- Working on a specific low-abundance target where a standard ELISA runs out of room? Talk to the Exazym Support Team → · support@cavidi.se