Abstract

BackgroundVarious chromatin modifications, identified in large-scale epigenomic analyses, are associated with distinct phenotypes of different cells and disease phases. To improve our understanding of these variations, many computational methods have been developed to discover novel sites and cell-specific chromatin modifications. Despite the availability of existing methods, there is still room for further improvement when they are applied to resolve the histone code hypothesis. Hence, we aim to investigate the development of a computational method to provide new insights into de novo combinatorial pattern discovery of chromatin modifications to characterize epigenetic variations in distinct phenotypes of different cells.ResultsWe report a new computational approach, ChARM (Combinatorial Chromatin Modification Patterns using Association Rule Mining), that can be employed for the discovery of de novo combinatorial patterns of differential chromatin modifications. We used ChARM to analyse chromatin modification data from the livers of normal (non-cancerous) mice and hepatitis B virus X (HBx)-transgenic mice with hepatocellular carcinoma, and discovered 2,409 association rules representing combinatorial chromatin modification patterns. Among these, the combination of three histone modifications, a loss of H3K4Me3 and gains of H3K27Me3 and H3K36Me3, was the most striking pattern associated with the cancer. This pattern was enriched in functional elements of the mouse genome such as promoters, coding exons and 5′UTR with high CpG content, and CpG islands. It also showed strong correlations with polymerase activity at promoters and DNA methylation levels at gene bodies. We found that 30 % of the genes associated with the pattern were differentially expressed in the HBx compared to the normal, and 78.9 % of these genes were down-regulated. The significant canonical pathways (Wnt/ß-catenin, cAMP, Ras, and Notch signalling) that were enriched in the pattern could account for the pathogenesis of HBx.Conclusions ChARM, an unsupervised method for discovering combinatorial chromatin modification patterns, can identify histone modifications that occur globally. ChARM provides a scalable framework that can easily be applied to find various levels of combination patterns, which should reflect a range of globally common to locally rare chromatin modifications.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-016-1307-z) contains supplementary material, which is available to authorized users.

Highlights

  • Various chromatin modifications, identified in large-scale epigenomic analyses, are associated with distinct phenotypes of different cells and disease phases

  • We report ChARM, a new computational approach based on association rule mining (ARM), which is de novo pattern discovery of differential chromatin modifications that occur globally in hepatocellular carcinoma (HCC) tissues of hepatitis B virus X (HBx)-transgenic mice

  • By applying ARM to three different types of histone lysine methylation, DNA methylation, and RNA polymerase II (Pol II) phosphorylation on a genomewide scale, we discovered 2,409 association rules that were expressed as combinatorial patterns of differential chromatin modifications

Read more

Summary

Introduction

Various chromatin modifications, identified in large-scale epigenomic analyses, are associated with distinct phenotypes of different cells and disease phases. To improve our understanding of these variations, many computational methods have been developed to discover novel sites and cell-specific chromatin modifications. We aim to investigate the development of a computational method to provide new insights into de novo combinatorial pattern discovery of chromatin modifications to characterize epigenetic variations in distinct phenotypes of different cells. Mis-regulation of gene expression, or attenuated post-translational modifications can impair the activity of histone-modifying enzyme complexes, and this may affect the mechanism that regulates global histone modifications throughout the genome. In this study, we aim to develop a new data mining approach for discovery and interpretation of differential GHMPs in cancer, and we use it to investigate whether an understanding of epigenetic alterations in cancer cells can expand prognostic capabilities

Objectives
Methods
Results
Discussion
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call