Symbolical Kernel-Based Reasoning: Its Application to the Rule Extraction in Dictyostelium discoideum DNA

Kernel-based method is one of the promising algorithms in the pattern recognition field [1]. We have extended the algorithm to the symbolical reasoning tasks (we call the algorithm symbolical kernel-based reasoning (SKBR)) [2], and we currently apply the SKBR to the rule extraction in Dictyostelium discoideum the DNA and cDNA databases of which are well established already [3]. Goal of our project is to show effectiveness of the SKBR in the genome informatics and to find out expression rules hidden in the DNA sequences, especially in the untranslated regions of Dictyostelium discoideum. We also go toward developing a new genome-rule-reasoning system including hardware for computational acceleration.