Aims: To determine the origins of DNA sequences isolated from the rumen microbial ecosystem using a self-organizing map (SOM). Methods and Results: DNA sequences other than 16S small subunit ribosomal RNA (SSU rRNA) gene sequences that were detected from the rumen were analysed by the SOM method reported by Abe et al. [2000, Self-Organizing Map (SOM) unveils and visualizes hidden sequence characteristics of a wide range of eukaryote genomes. Gene 365, 27-34]. Because query sequences positioned by SOM were scattered on the master drawing of SOM, it was suggested that many DNA sequences isolated from the rumen were collected from a broad range of micro-organisms. Although the results obtained by SOM were similar to those obtained by the neighbour-joining (NJ) method, SOM was able to presume the phylotypes of the query sequences without information about the 16S SSU rRNA gene sequences and homology searches, and to reveal existence of novel micro-organisms deduced to be cellulolytic bacteria, archaea and methanotrophic bacterium. Conclusions: As the SOM method defined phylotypes of unreported rumen micro-organisms, it is presumed that these phylotypes would be involved in rumen fermentation in cooperation with known rumen micro-organisms. Moreover, it is demonstrated that SOM is a useful tool for affiliating DNA sequences, which have no matches in databases. Significance and Impact of Study: Through SOM analysis, a better means of identifying rumen micro-organisms and estimating their roles in rumen function was provided.