Abstract
Identifying the sequence composition of different splicing modes is a challenge in current research. This study explored the dispersion distributions of 6-mer subsets in human acceptor splicing regions. Without differentiating acceptor splicing modes, obvious differences were observed across the upstream, core, and downstream regions of splicing sites for 16 dispersion distributions. These findings indicate that the dispersion value of each subset can effectively characterize the compositional properties of splicing sequences. When acceptor splicing sequences were classified into common, constitutive, and alternative modes, the differences in dispersion distributions for most of the XY1 6-mer subsets were significant among the three splicing modes. Furthermore, the alternative splicing mode was classified into normal, exonic, and intronic sub-modes, the differences in dispersion distributions for most of the XY1 6-mer subsets were also significant among the three splicing sub-modes. Our results indicate that dispersion values of XY1 6-mer subsets not only revealed the sequence composition patterns of acceptor splicing regions but also effectively identified the differences in base correlation among various acceptor splicing modes. Our research provides new insights into revealing and predicting different splicing modes.