Investigating the Intrinsic Differences in Flank Regions of Exon-Intron Junction Sites

The flank regions of exon-intron junction sites (EIJSs) are closely related to gene structure; and their sequence context is the determinants for the spliceosome to recognize the authentic EIJSs. Although some efforts had been devoted to mining sequence information in the flank regions of EIJSs, significant breakthroughs are still awaiting. In this paper, a hypothesis on EIJSs´ flank regions was proposed, it stated that the sequence context of intron flank is more flexible than the conjoined exon flank for a specific EIJS; it means that the exon flanks are more conservative than the intron flanks. For investigating the proposed hypothesis, a frameshift strategy was used to explore the sequence context; and a codon-based measurement, frameshift oscillation (FO), was devised to estimate the sensitivity of being disturbed by frame shifting; according to the hypothesis, the exon flanks will be more sensitive to frameshift than their conjoined intron flanks. After investigating all EIJSs in complete human genome with various lengths of flank regions by FO, the results reveal there do exist intrinsic differences in the flank regions of EIJSs. Therefore, the proposed hypothesis has great implications in predicting gene structure; furthermore, it is believed that the devised estimator FO really catches a key feature of authentic EIJSs and it is of great potential in identifying authentic EIJSs.