Histone modifications involved in cassette exon inclusions: a quantitative and interpretable analysis

Background: Chromatin structure and epigenetic modifications have been shown to involve in the co-transcriptional splicing of RNA precursors. In particular, some studies have suggested that some types of histone modifications (HMs) may participate in the alternative splicing and function as exon marks. However, most existing studies pay attention to the qualitative relationship between epigenetic modifications and exon inclusion. The quantitative analysis that reveals to what extent each type of epigenetic modification is responsible for exon inclusion is very helpful for us to understand the splicing process. Results: In this paper, we focus on the quantitative analysis of HMs' influence on the inclusion of cassette exons (CEs) into mature RNAs. With the high-throughput ChIP-seq and RNA-seq data obtained from ENCODE website, we modeled the association of HMs with CE inclusions by logistic regression whose coefficients are meaningful and interpretable for us to reveal the effect of each type of HM. Three type of HMs, H3K36me3, H3K9me3 and H4K20me1, were found to play major role in CE inclusions. HMs' effect on CE inclusions is conservative across cell types, and does not depend on the expression levels of the genes hosting CEs. HMs located in the flanking regions of CEs were also taken into account in our analysis, and HMs within bounded flanking regions were shown to affect moderately CE inclusions. Moreover, we also found that HMs on CEs whose length is approximately close to nucleosomal-DNA length affect greatly on CE inclusion. Conclusions: We suggested that a few types of HMs correlate closely to alternative splicing and perhaps function jointly with splicing machinery to regulate the inclusion level of exons. Our findings are helpful to understand HMs' effect on exon definition, as well as the mechanism of co-transcriptional splicing.