ELISA formats controls and readout

ELISA formats, controls and readout are the three decisions that turn an antibody binding event into a number you can say out loud. The plate does not know your hypothesis. It reports enzyme activity. You earn the right to call that activity a concentration by choosing a format that matches the analyte, by running controls that would have caught a sticky well, and by reading only inside a standard curve. Choosing the primary antibody itself is the companion guide choosing and checking a primary antibody. This page is a research explainer, not a diagnostic instruction.

Which format matches the analyte

In a direct assay the antigen is coated on the plastic and a labelled primary antibody binds it. It is simple and it needs a relatively pure antigen. An indirect assay uses an unlabelled primary antibody and a labelled secondary antibody, which adds signal and also adds a place for the secondary to stick. A sandwich assay coats a capture antibody, lets the analyte bind from a mixture, and detects it with a second antibody on a different epitope. That is the usual choice for a protein in serum, lysate, or culture medium, because the capture step supplies specificity the plastic cannot.

A competitive assay incubates the sample with a limited amount of antibody, or competes the sample against a labelled antigen. Signal falls as the analyte rises. Many small molecules are measured this way because they have only one epitope and cannot form a sandwich. Read the direction of the curve before you rank samples. A darker well is not automatically "more analyte".

The antibodies have to fit the format. A sandwich needs two clones, or a polyclonal pair, that can bind at once. Check that claim with a control antigen, not with hope. Sequence context for the target, including domains a capture peptide might have missed, belongs on a page such as UniProt.

How binding becomes a number

Most research plates use an enzyme conjugate, often horseradish peroxidase or alkaline phosphatase, and a substrate that becomes coloured or luminescent. The microplate reader in the photograph is the measurement, and it is only as meaningful as the wells you built. For a colorimetric substrate, the wavelength and any stop reagent belong to that substrate. Copy them from the reagent you opened. A kinetic read and an endpoint read are different experiments. Say which one you did.

The standard curve is a set of known concentrations taken through the same protocol. Fit a line only when the chemistry is actually linear. Many ELISA curves are sigmoid across a wide range, and a four-parameter fit is a common description. Choose the fit because it matches the curve, and lock that choice before you interpret outliers. Concentrations below the lowest standard or above the highest are outside the assay. Dilute or concentrate, or report them as out of range.

Matrix is part of the curve. A standard diluted in buffer can sit on a different line from the same standard diluted in serum. Spike a known amount into the real sample matrix and ask what fraction you recover. Poor recovery means the number is a rank at best. Dilutional parallelism, where a diluted sample falls on the same line, is the check that you are measuring the same thing at different doses.

Blocking, washes and the primary antibody

Blocking is a protein or detergent solution that covers bare plastic after the capture step. Without it, the detection antibody binds the plate and every well lights up. The no-analyte well is the proof. If that well matches your low samples, you do not have a low sample. You have a background.

Washes remove unbound reagent. Too few washes raise the background. Too many, or a harsh buffer, can strip a weak interaction. Record the buffer and the count. The primary antibody's dilution is empirical. A checkerboard of capture and detection dilutions, run once when the assay is set up, saves every later plate. The antibody guide explains how to decide the reagent is specific before you invest in that checkerboard.

Method collections such as protocols.io show how other laboratories structure a plate map. Keep your map in the same file as the absorbance values. A plate map reconstructed the next day is how standards and samples trade names.

Controls worth the wells

A substrate blank contains substrate and stop, without enzyme. It catches a substrate that has already gone off. A zero standard contains matrix and no added analyte. It catches background and contaminated diluent. A positive control of known concentration, ideally a lot you kept for this purpose, catches a dead enzyme or a forgotten antibody. A well with no capture antibody, or no primary antibody, catches reagents that bind plastic or the sample directly.

Replicates show pipetting. They do not create biological replication. If you plated one culture three times, you have one biological result.

Failure modes

The hook effect, already noted, is the dramatic one. Edge effects are the quiet one. Outer wells evaporate faster, concentrate the reagent, and drift. If the plate map puts all the controls in the centre and all the treated samples on the edge, the map is the result. Rotate layouts, or avoid edges, and seal during incubations. Temperature gradients in a cold room or a sunny bench do the same thing.

A biotin-avidin amplification system fails in samples rich in biotin, including some milk blockers. A phospho-specific detection fails in milk that contains phosphoproteins. Cross-reactivity of the primary antibody is a reagent problem, taken up in the antibody article. An expired substrate turns the whole plate yellow, including the blank. That plate is not "high expression".

Safety and research limits

Antibodies and standards may contain azide or other preservatives. Stop solutions are often acid. Substrates can be irritants. Human serum and culture samples that might contain pathogens follow institutional biosafety rules. The WHO Laboratory Biosafety Manual and the CDC BMBL are public references for that institutional decision. They are not a permit, and this article is not a diagnostic approval.

Heat, humidity and the reader

Colour development is a timed enzyme reaction. A hot afternoon shortens it. Keep the development in the temperature window the substrate states, and use the same timer for every plate you will compare. Humidity and a loose lid change edge wells first. If a power cut stops the reader after development, the colour may still be on the plate. Read it promptly and record the delay. A stopped reaction is more patient than a kinetic one. Do not leave a developed plate in direct sun while you find a spare reader.

What to put in an enquiry

Name the analyte, the sample matrix, the species, whether you need a sandwich or a competitive format, the concentration range you care about, and any biotin or phospho constraint. The reagents and chemicals catalogue is the sourcing start for buffers, plates and related reagent classes. The molecular biology pathway is the wider methods context. Use the quote request to ask whether a quotation is possible. Do not treat a catalogue line as a coated plate that is already validated for your matrix, and do not read this site as operating a diagnostic ELISA service. Ask for the antibody pair and the curve range in the reply.