The literature concerning the subcellular location of Y-box binding protein 1 (YB-1), its abundance in normal and cancer tissues, and its prognostic significance is replete with inconsistencies. An explanation for this could be due in part to the use of different antibodies in immunohistochemical and immunofluorescent labeling of cells and tissues. The inconsistencies could also be due to poor resolution of immunohistochemical data. We analyzed two cohorts of breast tumours for both abundance and subcellular location of YB-1 using three different antibodies; two targeting N-terminal epitopes (AB-a and AB-b) and another (AB-c) targeting a C-terminal epitope. We also investigated stress-induced nuclear translocation of YB-1 in cell culture. We report that both AB-a and AB-c detected increased YB-1 in the cytoplasm of high-grade breast cancers, and in those lacking estrogen and progesterone receptors; however the amount of YB-1 detected by AB-a in these cancers is significantly greater than that detected by AB-c. We confirm our previously published findings that AB-b is also detecting hnRNP A1, and cannot therefore be used to reliably detect YB-1 by immunohistochemistry. We also report that AB-a detected nuclear YB-1 in some tumour tissues and stress treated cells, whereas AB-c did not. To understand this, cancer cell lines were analyzed using native gel electrophoresis, which revealed that the antibodies detect different complexes in which YB-1 is a component. Our data suggest that different YB-1 antibodies show different staining patterns that are determined by the accessibility of epitopes, and this depends on the nature of the YB-1 complexes. It is important therefore to standardize the protocols if YB-1 is to be used reproducibly as a prognostic guide for different cancers.