Epigenetic methodologies for behavioral scientists

Abstract

Hormones are essential regulators of many behaviors. Steroids bind either to nuclear or membrane receptors while peptides primarily act via membrane receptors. After a ligand binds, the conformational change in the receptor initiates changes in cell signaling cascades (membrane receptors) or direct alternations in DNA transcription (steroid receptors). Changes in gene transcription that result are responsible for protein production and ultimately behavioral modifications. A significant part of how hormones affect DNA transcription is via epigenetic modifications of DNA and/or the chromatin in which it is entwined. These alterations lead to transcriptional changes that ultimately define the phenotype and function of a given cell. Importantly we now know that environmental stimuli influence epigenetic marks, which in the context of neuroendocrinology can lead to behavioral changes. Importantly tracking epigenetic states and profiling the epigenome within cells require the use of epigenetic methodologies and subsequent data analysis. Here we describe the techniques of particular importance in the mapping of DNA methylation, histone modifications and occupancy of chromatin bound effector proteins that regulate gene expression. For researchers wanting to move into these levels of analysis we discuss the application of modern sequencing technologies applied in assays such as chromatin immunoprecipitation and the bioinformatics analysis involved in the rich datasets generated.

Figure 1

Summary of tools for low-level analysis (both Illumina software and 3rd party tools) as they apply to the Illumina Genome Analyzer Pipeline.

Figure 2

Plot of H3K9/14Ac and H3K4me3 average read number relative to the TSS start site of genes stratified into 5 bins according to gene expression ranging from the lowest (0–20%) to highest (>80%): sort the genes on the Affymetrix U133 array according to their expression level and put the lowest 20% in one bin, the next 20–40% in another and continue until we have 5 bins with the same number of genes representing low to high expressed genes. For each group of genes separately, we calculate the number of mapped reads at every genomic coordinate (ranging from −2kb to +2kb) relative to the TSS of every gene in the group.

Figure 3

Boxplots of log2 gene expression levels for genes with no modifications, H3K4me3, H3K9/14Ac and both H3K4me3 and H3K9/14Ac within 1kb of their 5' ends. We classify a gene into one of these four categories based on the present or absence of the H3K4me3 and H3K9/14Ac sites within 1kb of their 5' ends. The lower part of box is 25th percentile; the thick middle line of box is the median; the upper part of the box is 75^th percentile; and the lower and upper whiskers are 5th and 95th percentile and represent outliers.