A 613-bp fragment of an essential ketosynthase gene from the biosynthetic pathway of aromatic polyketide antibiotics was sequenced from 99 actinomycetes isolated from soil. Phylogenetic analysis showed that the isolates clustered into clades that correspond to the various classes of aromatic polyketides. Additionally, sequencing of a 120-bp fragment from the ␥-variable region of 16S ribosomal DNA (rDNA) and subsequent comparative sequence analysis revealed incongruity between the ketosynthase and 16S rDNA phylogenetic trees, which strongly suggests that there has been horizontal transfer of aromatic polyketide biosynthesis genes. The results show that the ketosynthase tree could be used for DNA fingerprinting of secondary metabolites and for screening interesting aromatic polyketide biosynthesis genes. Furthermore, the movement of the ketosynthase genes suggests that traditional marker molecules like 16S rDNA give misleading information about the biosynthesis potential of aromatic polyketides, and thus only molecules that are directly involved in the biosynthesis of secondary metabolites can be used to gain information about the biodiversity of antibiotic production in different actinomycetes.Soil actinomycetes, especially those that belong to the genus Streptomyces, have been the focus of intensive research for the past several decades. The interest in Streptomyces arose from the finding that this group of bacteria seems to have the ability to produce a large variety of different bioactive compounds that have a wide spectrum of activity. From the 1950s to the mid-1970s numerous new bioactive molecules were discovered through large screening programs, and these molecules subsequently found their way into various clinical uses ranging from control of infections to cancer treatment (21).In more recent years, modern high-throughput screening methods have exponentially increased the number of strains screened annually, but the number of novel compounds discovered has not increased in the same proportion. One of the many reasons, presumably the most important one, for this problem is that old molecules (and strains) are being rediscovered with the screening procedures that are in use today (21).In a previous paper (18) Metsä-Ketelä et al. reported a method that could be used for preliminary classification of strains on the basis of their genetic abilities to produce various compounds belonging to the aromatic polyketide group. This method is based on PCR amplification of a gene fragment that is essential in the biosynthesis pathway of aromatic polyketides and on analysis of the amplified regions by phylogenetic methods. The degenerate primers designed for this purpose amplify a portion of a ketosynthase gene (KS ␣ ), which in collaboration with KS  and an acyl carrier protein condenses small carboxylic acids in a stepwise manner to form a long polyketide chain that is subsequently folded into a range of different aromatic compounds by various ketoreductases, cyclases, and aromatases. Later, the molecule formed is of...