Large-scale information entropy analysis of important sites in mature and precursor miRNA sequences.

Abstract

In order to find evidence of consistent sequence conservation or the base correlation degree in miRNA, some important sites in the sequences of reported miRNA and their precursors (pre-miRNA) were investigated via information entropy analysis. Twelve different groups of sites were obtained from special locations (head, tail) in miRNAs of different sources according to taxonomy (animal, plant and virus) and then analyzed by measuring the single base information redundancy (D(1)(L)) and the adjacent base related information redundancy (D(2)(L)). The results showed that D(2)(L) has more information than D(1)(L), though D(1)(L) changes roughly consistently with D(2)(L) in each group. Viral pre-miRNAs are more conservative than those belonging to animals or plants. In addition, U is dominant in most sites compared with other nucleotides. It was also found that in the middle of several groups, there were sites where miRNAs were cut down from pre-miRNAs by Enzyme Dicer which were significantly conservative. This phenomenon shows that the conservatism is an aspect of the of miRNA and may be involved in the recognition and cutting by the Dicer. Those results provided another perspective for understanding more about the primary structure of pre-miRNA.