Western blot as an antibody specificity check

Specificity, in this argument, means the signal depends on the protein you named and not on a neighbour that happened to be in the lane. Cross-reactivity means the antibody also binds a protein you did not name. A western blot is one place those definitions become visible, because the gel spreads proteins by apparent mass before the antibody ever sees them. An ELISA well does not offer that spread. The blot is therefore a useful suspicion test before a plate, and a poor certificate if you stop at a single pretty band. How the membrane is made is the companion western blot from gel to membrane. How to read the gel that preceded it is reading a protein gel. Plate context lives in ELISA formats, controls and readout. The header photograph is a membrane being lifted from a tray. This page is about what the bands allow you to say.

A short glossary for the claim you want to make

The ladder, or molecular-weight marker, is a set of proteins of known mass run in a neighbouring lane. It lets you estimate relative molecular mass. It is not a standard curve for quantity, and it is not proof you loaded the sample you think you loaded.

The membrane is the nitrocellulose or PVDF sheet that now holds the proteins. Antibodies probe the membrane, not the gel. A transfer that failed leaves a gel full of protein and a membrane that stays blank for a boring reason.

Blocking is the protein or detergent step that occupies bare membrane. Without it the antibody sticks everywhere and the lane is a smear of chemistry rather than of biology. Milk, albumin and casein are the usual classes, with the same biotin and phospho-protein caveats as on a plate.

Exposure is the interval during which enzyme and substrate produce light, colour or fluorescence. A short exposure can hide a weak true band. A long exposure can promote a faint cross-reactive band into the figure. Exposure is a display choice. It is not specificity.

Apparent mass is where the band sits relative to the ladder. Expected mass is the mass you predicted from the sequence, plus any modification you have a reason to expect. The two can differ for honest reasons: glycosylation, cleavage, an unusual gel composition. They can also differ because the antibody bound something else.

What one band at the right mass is allowed to mean

Load a lysate you expect to contain the target. Probe with the antibody. See one band near the predicted mass. You may say the antibody can recognise a protein of that mass in that lysate, and that it does not paint the rest of that particular blot under that exposure. You may not yet say the band is the target. A different protein of the same apparent mass would look identical. You may not say the antibody will be specific in a sandwich, where that neighbour would contribute signal without a gel to unmask it.

Look up the predicted mass and the known relatives on UniProt before you decide the band is "about right". A 15 percent miss that you shrug at may be a different protein entirely. Background on antibody detection as a molecular method is sketched on the NCBI Bookshelf. The local lanes still do the work.

The lanes that make the sentence stronger

A knockout or a knockdown lane is the cleanest genetic control. The same cells, the same load, the target reduced or absent, the band reduced or absent with it. A band that stays after a verified knockout is not the target, however perfect its mass. A band that vanishes while a loading control stays is the observation you wanted.

A peptide-competition lane is the next best when you have no genetic null. Incubate the antibody with the immunising peptide, or with the recombinant antigen, before it meets the membrane. Specific binding is absorbed. Bands that remain are sticking by some other route. Use a control peptide that should not compete, if you have one, so you can see that any peptide at that concentration does not simply poison the antibody.

A recombinant protein lane shows that the antibody can bind the intended sequence when it is undeniably present. It does not show that the lysate band is the same protein. Run it as an extra lane, not as a substitute for the lysate.

An antibody-only lane, more precisely a lane with no primary antibody, shows what the secondary and the detection reagents do to that membrane. A band there belongs to the secondary, often to immunoglobulin in the sample. It must not be counted as target.

A second lysate that should lack the target for a biological reason, such as a cell line with independent evidence of no expression, is weaker than a knockout and still better than a single lane. Cell-line identity is its own problem. Do not build the negative lane on a misnamed line.

Method habits for these lanes are widely written up, including on protocols.io. Detection chemistry timing is described by reagent makers, for example in Promega protocols. Follow the sheet of the substrate you opened.

A smear is data

Many bands, or a smear from top to bottom, means this antibody sees a crowd under these conditions. Diluting the antibody sometimes clears a dirty blot, and sometimes it simply fades the crowd along with the true band. A longer exposure is the wrong direction. If the ELISA you planned uses this antibody as capture or detection, the well will sum every protein the antibody sees. The plate cannot show you the ladder of bands you just ignored. Either map which band depends on the target, or choose a different reagent before you invest in a standard curve.

How this changes the ELISA you run next

If the blot is clean in the genetic sense, you have a reason to try the antibody in a sandwich or as a coat, and you still repeat a negative sample on the plate. The well has no ladder. If the blot is crowded, do not move that antibody into an ELISA and hope quantification will average the doubt away. A precise standard curve of a cross-reactive antibody is a precise measurement of a mixture.

If the blot band and the ELISA disagree, believe the disagreement. Denatured epitope and native epitope are different surfaces. The antibody guide is the place to retire a reagent. This glossary only fixes what the blot words mean so the retirement is based on a sentence you can defend.

Transfer chemistry and research limits

Methanol, transfer buffers, acrylamide from the gel that preceded the blot, and chemiluminescent substrates each have handling rules set by the institution. The membrane in the photograph is being moved with forceps for a reason. Gloves and eye protection follow the local assessment. Nothing in a specific-looking band makes the method diagnostic. A research blot supports a research statement about the samples in the lanes.

When the power fails mid-transfer

A transfer is a timed electrical step. If the supply drops halfway, proteins have moved some unknown distance. Do not probe that membrane and interpret a weak band as low expression. Repeat the transfer. In a warm laboratory a tank that was already heating can cook a gel during a long transfer. That smear is heat damage. Shorten the time or improve cooling as the protocol allows, and do not lengthen the exposure to compensate. Record the power interruption next to the notebook page so a missing band is not later called biology.

What to ask when you need an antibody you can blot

Name the target accession, the predicted mass, the species of the lysate, and whether you have a knockout or only a peptide for competition. Say that you want the reagent for a specificity check and, separately, whether you also need it for ELISA. The two claims should not be sold as one. The reagents and chemicals catalogue covers related reagent classes. The molecular biology pathway is the methods setting. A blot-capable antibody choice can be discussed through the quote request. Ask which epitope the reagent sees, because a denatured-linear epitope and a native epitope support different assays.