It is not uncommon that, contrary to the theoretical predictions, several bands are detected. Although it is possible that the antibody is not entirely specific for the protein, other factors may be responsible:

• Proteolytic breakdown of the antigen. This is not uncommon, particularly if samples are stored for prolonged time or if proteins or membranes are fractionated after homogenization of the starting tissue. All additional bands are of lower apparent molecular mass than the full-length protein. Particularly susceptible are synapsins and synaptotagmins. Addition of protease inhibitors such as PMSF, pepstatin or leupeptin should be considered. Several vendors offer protease cocktail reagents for this purpose. It is possible, of course, that some of the protein was degraded prior to extraction, in which case it is not possible to remove these products.
• Insufficient denaturation and/or reduction. If the sample buffer does not contain sufficient SDS and/or reducing agent (DTT, 2-mercaptoethanol), the protein may not be fully dissociated into its subunits, reduced or denatured. Hence, extra bands may appear above the desired protein on the gel. Boiling the sample in sample buffer immediately prior to loading can reduce these noncovalent interactions or disulfide linkages. Use of an irreversible reducing agent such as TCEP, instead of DTT or beta-mercaptoethanol, may also be helpful.
• Too much protein per lane or detection system too sensitive. Overloading of the gel is one of the most common reasons for "ghost bands". Immobilized proteins may provide a concentrated adsorptive surface to which certain IgG may bind nonspecifically. Similarly, such nonspecific binding may be uncovered when highly sensitive detection systems such as enhanced chemiluminescence are employed. A dilution series of the starting material usually clarifies which of the signals are artefactual.
• Inefficient blocking. A variety of different blocking agents are described in the literature including nonionic detergents and/or proteins. Change of the blocking conditions may remedy the problem.
• Concentration of antigen too low. The resolution of SDS-PAGE is limited to 50-100 bands. If the relative concentration of the antigen of interest is too low (less than 0.2% of total protein), it may be difficult to detect (for instance, synaptobrevin/VAMP comigrates with histones in cell homogenates which interfere with its detection). Signal enhancement may then lead to the appearance of artefactual bands. Enrichment of the antigen by fractionation or by immunoprecipitation should be considered.
• Unwanted specific reactivities. Antibodies, especially polyclonal antibodies, will sometimes recognize proteins in a biological sample with similar sequences as the antigen, such as those in the same protein family. Non-affinity-purified polyclonal antibodies will often recognize other bands due to the exposure of the animal to various other proteins in its lifetime. This can be addressed by requesting pre-immune serum and using this as a comparison.