Mapping the Genome Sequences on Self Organizing Maps

Recently, all genome of some species were determined and the functions of many genes were also determined. But, it will need much efforts to determine the functions of all genes. Furthermore, not all regions (regulatory, promoters, exons, introns and etc.) on the genome sequences are distinguished yet. To determine the functions of the genes or the regions on genome, the comparison between the target gene and known genes is often made using the alignments between them. But, the amounts of computations for examining the alignments between target gene and many known genes are very large. Thus, the method which can efficiently compare the sequences is desired. Self Organizing Map(SOM)s [3] were proposed by Kohonen and widely used for clustering, classifying and visualizing the multi-dimensional data. SOMs were also used in bio-informatics for visualizing the gene expression data. The genome sequences of some species were also visualized on 2-dimensional SOM maps according to the frequency of n-tuples of nucleotides (A, G, T, C) [1]. We are developing the modified SOM algorithm(Sequence SOM: SSOM) for designing the DNA chips which can efficiently organize the features of the DNA sequences [2]. In this paper, we used the map organized by SSOM for mapping the genome sequences on 2-dimensional planes and the results of the comparison of genes taken from different pathways are shown.