The human milk fat globule has proved to be a good source of antigenic material for production of antibodies against surface components of breast epithelial cells. Monoclonal antibodies against one of the major components of the human milk fat globule, which identify a glycoprotein with an apparent molecular weight of 46,000, have been found to be useful for both breast cancer diagnosis and therapy. In order to characterize this Mr 46,000 glycoprotein, specific monoclonal antibodies were used to select complementary DNAs from a lambda gt11 expression library from lactating breast. The largest complementary DNA insert (BA46-1) was 1270 base pairs and encoded 217 amino acids. A single 2.2-kilobase RNA was specifically detected in a variety of carcinoma cell lines, using this complementary DNA probe, and it was overexpressed in some carcinoma lines. The mRNA levels correlated with the level of expression of the antigen in these cell lines as detected by Western blot analysis. Sequence analysis revealed strong homology of the Mr 46,000 glycoprotein with serum factors VIII and V, in the region implicated in phospholipid binding.