Epidemiological studies of Streptococcus agalactiae strains have been limited by the lack of sensitive and discriminatory methods for comparing clinical isolates. Serotyping, albeit a widely used methodology, has been shown to possess low capability to distinguish between epidemiologically related and unrelated isolates. We have employed here a random amplification of polymorphic DNA (RAPD) assay, using degenerate oligonucleotides as primers, to characterize S. agalactiae isolates from related or unrelated clinical samples. Epidemiologically-related isolates (mother-infant pairs) showed identical profiles by this methodology. On the contrary, 12 epidemiologically-unrelated isolates (classified into 5 different serotypes) resulted in 11 distinct RAPD patterns. This suggests that the proposed modified RAPD assay provides a highly discriminatory tool for the analysis of genomic diversity among isolates from pathogenic organisms.