Based on the characteristic of nucleotide distribution in nucleosome positioning and inhibiting sequences, the method of Increment of Diversity with Quadratic Discriminant (IDQD) was applied to the classification of these two types of sequences. The mean area under ROC curve archives 0.958. By using this model, the nucleosome formation potential was analyzed in the regions around the splice sites (GT/AG). The results show that coding regions have a high potential to form the nucleosome and the primary RNA transcripts are rigid, while DNA sequences corresponding to the splice sites and their adjacent intron regions tend to be nucleosome free and the primary transcripts from these regions are relative flexible. Moreover, the negative correlation between nucleosome positioning/inhibiting of DNA sequences and RNA flexibility/rigidity is demonstrated around the splice sites, providing a mechanism for understanding the correlation between the nucleosome positioning of DNA and the splicing of transcribed RNA sequences.

Based on the conservation of nucleotides at splice sites, the characteristics of base composition and base correlation in the adjacent segment sequences, the distance between alternative donor or acceptor splice sites and the content of GC and TC near splice sites, the donor and acceptor splice sites for alternative and constitutive introns are predicted by use of the method of diversity measure combined with quadratic discriminant analysis. For alternative splice sites the total prediction accuracies are 87.9% and 89.9% for donors and acceptors respectively (with the chosen threshold -2). For constitutive splice sites the total accuracy are 92.8% and 94.3% for donors and acceptors respectively (with the chosen threshold -1).