Typing of exon-skipping events based on digitized methylation signals using a DTW-based FSOM (Part of the introduction)

A flexible weighted vector adjustment based Self-Organizing-Map (FSOM) published by [1] was developed for but not limited to methylome analysis. One of its main purposes is to study the correlation between AS events and DNA methylation patterns, but besides that, the FSOM can be used for any kind of digitized epigentic signals. Therefore, it utilizes the Dynamic-Time-Warping (DTW) concept for the comparison of such data. It has already been tested on methylation data from eight different human tissues and certain patterns have been found and could also be associated with other genomic propertes. These studies are also made possible by a broad range of freely available NGS data experimentally generated from (epi-)genome-wide analyzes of different tissues/cell types. In the context of this written elaboration, the FSOM will be used for the typing of AS events based on methylation data obtained from human spleen cells. To provide some biological background, the introductory chapter will at first cover some general aspects about epigenetics.

Introduction

A general definition for epigenetics does not exist, since the term is used in various ways. Nevertheless, a lot of descriptions have in common, that the object of study is mainly the transmittance of gene-regulatory information over several generations of cells or species. As the prefix “epi” already indicates, changes in the underlying (largely static) genome are not involved. This is in accordance with a widely accepted definition given by Riggs et al. in 1996, which states that epigenetics is “the study of mitotically and/or meiotically heritable changes in gene function that cannot be explained by changes in DNA sequence”. In the course of reproduction, chemical alterations within the epigenome are passed down to an organisms offspring. During the fertilization process, gametes still carrying parental epigenetic information fuse together and thereby form a single-celled zygote from which a new individual can arise. After a fundamental, but still incomplete reprogramming of the zygote into a pluripotent state, some of the meiotically inherited marks still remain even in very early embryonic stages. Pluripotent embryonic stem cells can then differentiate into several somatic cell types, which ultimately results in the development of functional tissues and organs. Throughout morphogenesis, progressive changes in the epigenetic code lead to activation of particular differentiation-associated genes as well as inactivation of genes that are crucial for pluripotency. These modifications are mitotically propagated to upcoming cell generations causing their gene expression patterns to be programmed into a more defined and restricted way. In eukaryotic cells genetic information is contained within chromatin, which mainly consists of nucleic acids and proteins like histones. Epigenetic changes affecting the overall structure of chromatin are chemical modifications of these core components. These variations can either have activating or repressing effects leading to eu-chromatin or hetero-chromatin respectively. Members from different histone families act as subunits, that form octameric protein complexes. DNA is “wrapped around” these complexes thereby creating nucleosomes, which are the basic DNA packaging unit. Alterations like methylation or acetylation of their N-terminal tails influence their tertiary structure and lead to allosteric effects within the quartenary structure of such a multi-histone complex. Thus, the accessibility of genes to the transcription machinery is directly affected by their conformation. In early phases of cell development, genes required later are mainly repressed by highly flexible histone modifications, which is assumed to be a short term effect. In contrast to that, DNA methylation is considered to be a long term mechanism. Besides gene silencing, it is important for many other processes including transposon silencing, imprinting and X-inactivation. In mammals, the methylation of cytosines occurs primarily in the CpG context, but a small fraction appears to be methylated in a non-CpG context, whereby among the three other possible dinucleotides (e.g. CpA, CpC, CpT), CpA seems to be predominant. The highest concentration of CpG dinucleotides is generally observed around promoter regions, where they occur in so called CpG islands. Genes with hypomethylated promoter regions and hypermethylated gene bodies are thought to be actively expressed. During development however, methylation patterns can change. Promoters of genes, that are unnecessary in later cell stages, are actively methylated, which may inhibit the binding of transcription factors to certain promotor elements or result in the recruitment of methyl-CpG-binding repressors. In addition to that, hypermethylation seems to cause locally higher nucleosome occupancy, thereby also affecting DNA accessibility. Since methylation patterns of such promoter regions are stably propagated during mitotic cell divisions, the resulting inactivation proceeds in the very same manner in daughter cells. DNA methylation is thus an important epigenetic modification to achieve permanent silencing of certain genes in the course of cell specialization. The chemical modification of cytosines is done by DNA methyl transferases (DNMT) and their general function is the replacement of a hydrogen atom with a methyl group at C5-position, thereby turning it into a 5-methylcytosine. Members of the DNMT3 family are responsible for the de novo establishment of DNA methylation patterns. The DNMT1 family ensures a stable propagation of these patterns over multiple cell divisions by re-establishing the methylation marks on a newly synthesized strand of hemi-methylated DNA. Nevertheless, gene silencing accomplished by DNA methylation is not irreversible at all. During epigenetic reprogramming of primordial germ cells or of the earlier mentioned zygote, parental methylation marks must be actively removed to ensure the re-activation of pluripotency genes. The mechanism of active demethylation and its regulation are not fully unveiled yet.