A marker read in tissue is judged by its pattern in a known positive and a known negative on the same slide, not by a band on a blot, and the processing decides whether the pattern appears at all. Panels rather than single stains carry the conclusion, because almost every antigen in this class appears somewhere else in the body.
An epcam antibody and an ncam antibody on epithelial and neural antigens
The epithelial adhesion molecule is a lineage marker and an enrichment handle, and clones differ in whether they bind the extracellular domain that survives fixation or a region that needs retrieval, which is why a clone used for capture is not the clone used on a section. The neural adhesion molecule is expressed in neuroendocrine tissue and in natural killer cells, so its presence is informative only inside a panel, and its heavily glycosylated form migrates and stains differently between tissues.
A muc5ac antibody and a filaggrin antibody on secreted and cornified products
Mucin epitopes are glycoform dependent, so two clones against the same mucin can mark different cells in one section, and the glycosylation changes with disease, which is the point of measuring it and the reason a clone is cited rather than the protein. Filaggrin is processed from a large precursor in the granular layer, so a clone's epitope decides whether the precursor, the monomers or both are seen, and a normal skin control shows the expected layer-restricted pattern.
A collagen antibody and a collagen type i antibody, and which chain
Collagen antibodies have to name the type and often the chain, because the family shares sequence and structure, and they have to state whether the epitope is on the triple helix or on a denatured fragment, since retrieval and digestion change which is present. In tissue the stain is fibrillar and extracellular, which is the control that says it worked; in a lysate only the unassembled pool is present, so a blot of cells measures synthesis rather than matrix.
A claudin-5 antibody and an icam-1 antibody on endothelium
Claudin 5 is the tight junction protein that makes the blood brain barrier tight, so the informative readout is a junctional line between endothelial cells rather than an intensity, and its loss is read as barrier disruption. The adhesion molecule is induced by inflammatory stimuli, so an untreated section is the baseline and an induced one the positive control. Both are membrane proteins whose epitopes suffer under aggressive retrieval, which is where a weak stain usually comes from.
Why a panel rather than a single marker
Squamous, adeno and neuroendocrine lineages are separated by combinations rather than by one antigen, which is why diagnostic practice uses panels with expected positive and negative results for each candidate. The same logic applies in research: a claim about a cell type needs a lineage marker plus a functional one, imaged in channels that do not bleed, with the negative population shown rather than described. Panel design starts from what has to be excluded.
Fixation and retrieval decide most failures
Formalin crosslinks and masks epitopes, and heat retrieval in citrate or EDTA recovers them unevenly, so a clone validated on frozen sections is not validated on paraffin and a buffer change alters the result. Over-fixation is partly recoverable by longer retrieval; under-fixation loses morphology for good. The practical route is a fixation and retrieval matrix run once on control tissue for each new clone, with the winning combination recorded beside the clone and lot.
Controls on every run, not once
Three controls make a stain readable and they belong on every run rather than in a validation folder: a tissue known to be positive, a tissue or region known to be negative, and a no-primary section that shows what the detection system does on its own. For a marker supporting a diagnostic claim, an internal control within the same section is better still, since it experienced the same fixation and retrieval as the cells in question.
What to ask a supplier for this class
Ask for the clone identifier, the immunogen, and the vendor's own image in the tissue and the processing you will use, with the retrieval named. Ask what the expected pattern is, because that is what you will check against, and whether the clone is offered in a diagnostic grade if the result will support a clinical decision. Finally ask about lot size and availability, since a clone that disappears mid-study costs a revalidation.
Questions people ask about p40 antibody
Why do two clones against one mucin stain different cells?
Because mucin epitopes are glycoform dependent and the glycosylation changes with disease. The clone belongs in the method.
Can a frozen-section clone be used on paraffin?
Not on that evidence. Formalin crosslinking and heat retrieval change the epitope, so the combination has to be validated.
Which controls belong on every run?
A known positive, a known negative and a no-primary section, with an internal control in the same section where a diagnostic claim rests on it.