We report the identification of a novel cDNA representing an mRNA showing significantly higher levels of expression in benign breast lesions than in carcinomas. This cDNA was identified by differential screening of a cDNA library generated from a breast carcinoma, and shows consistently higher expression in fibroadenomas than in carcinomas. The expression in both benign and malignant tissues is highest in epithelial cells as determined by in situ hybridization to tissue sections. The nucleotide sequence of the full-length cDNA has been determined, and the deduced protein is highly basic with no signal or transmembrane sequence, but two potential nuclear localization signals. Neither the DNA nor the protein sequence show any significant homology to sequences in current databases. The cDNA hybridizes to multiple sequences within both human and other mammalian genomes, but to single genomic sequences in Drosophila, Physarum and Schizosaccharomyces pombe. This cDNA therefore represents a highly conserved gene sequence. We have identified only one major transcript in human cells, and it seems likely that there are several pseudogenes within the human genome.