Application of Immune Classifier Based on Increment of Diversity in the Model Species Genomes Identification

The model species genomes project has played a very important role in the research of human genomes. By comparing and identifying the genomes´ information in the different biological evolution stages with the model species genomes, it is favorable to understand deeply the genomes´ structure and function of the advanced species, especially the human being, and to reveal the life´s essential law. In the paper, the DNA sequences of three model species including C. elegans, S. cerevisiae and A.thaliana are divided into three types: intron, exon, intergenic DNA at first. The bases of each sequence are given number (1, 2, hellip, N) base by base from the beginning. According to three phases of the codons, we can extract three subsequences which numbers are respectively 3n+1, 3n+2 and 3n+3 (n=0, 1, 2, ...,N/3-1). Calculating respectively the probability of four bases in three subsequence we can attain 12 parameters which are regarded as the state parameters of the diversity sources. Each sequence of intron, exon and intergenic DNA is expressed by a 12-dimension eigenvector. All of these sequences´ 12-dimension samples are divided into the training sample set and the testing sample set. Then, according to the immune network theory, a immune classifier based on the diversity increment is constructed, which considers the reciprocal of the diversity increment as affinity function and makes the immune network evolve in the direction of identifying the antigens by submitting unceasingly the antigens one by one, which is a record of the training sample set. Finally, the classifier has high performance and its prediction accuracy is up to 85% by test.