Comprehensive Study Notes on Epigenomics: DNA Methylation, Histone Modifications, and Disease

Administrative Requirements for Proteomics Presentations

• Student Roster and Groups

  • There are a total of $54$ students enrolled in the course, divided into groups of three.
  • The Group ID is the identifying number located in the last column of the student table (highlighted in green).
  • Group IDs are used to determine which specific date a group will present and which dates they must attend the lectures.
  • Students are encouraged to reach out to their group members for planning, review, and study of course material.
  • Contact Information: If a group member's email is missing from the provided table, students should contact the professor for that information.

• Presentation Guidelines and Grading

  • Topics are assigned according to the Group ID.
  • Requirements include specific limits on the number of slides and the total duration of the presentation.
  • Grading Criteria:
    • The overall quality of the presentation.
    • Successful management of audience questions.
    • Timing: Groups must neither significantly exceed nor under-utilize their assigned time slots.
  • Submission Deadline: All slides must be submitted for every group by noon on May 21.

• Question Submission Assignment

  • Every student is required to submit several questions for each topic discussed.
  • Students must generate questions for their own assigned topic as well as for other topics.
  • These questions contribute a small percentage to the overall presentation grade.
  • Detailed instructions regarding the submission of slides and questions will be provided via separate communications.

Overview of Bisulfite Sequencing for DNA Methylation

• Definition and Function

  • Bisulfite sequencing is a method designed to detect DNA methylation patterns on a genome-wide scale.
  • It relies on bisulfite treatment, which is a chemical modification process.

• The Chemical Principle of Bisulfite Treatment

  • In theory, the treatment converts all cytosine ($C$) nucleotides to uracil ($U$).
  • Methylated cytosines are chemically resistant to this process and remain as cytosines ($C$).
  • This biochemical difference allows researchers to distinguish between methylated and unmethylated sites.

• Experimental Readout and Sequencing

  • Unmethylated Alleles: The $C$ is converted to $U$. During Polymerase Chain Reaction (PCR) and subsequent sequencing, the $U$ is read as a thymine ($T$). Therefore, an original unmethylated $C$ appears as a $T$ in the final sequencing readout.
  • Methylated Alleles: The treatment has no effect on methylated cytosines. The nucleotide stays as a $C$ through PCR and sequencing.
  • Conversion Summary: $C \rightarrow U \rightarrow T$ (unmethylated) vs. $C \rightarrow C$ (methylated).

Bioinformatics of Bisulfite Sequencing Alignment

• The Mismatch Abundance Problem

  • The first step in data analysis is aligning sequencing reads to a reference genome, similar to RNA-seq.
  • Standard aligners struggle with bisulfite reads because the conversion of unmethylated $C$ to $T$ creates massive mismatches relative to the reference genome.
  • Aligners typically seek to minimize differences; the abundance of $C$-to-$T$ transitions prevents aligners from effectively finding the correct genomic position.

• Reduced Complexity Mapping (Three-Nucleotide Genome)

  • To solve the alignment problem, researchers reduce the complexity of the genome and the reads.
  • Step 1: All genomic cytosines ($C$) in the reference genome are computationally converted to thymines ($T$).
  • Step 2: All cytosines ($C$) in the sequencing reads are also converted to thymines ($T$).
  • Result: This creates a "three-nucleotide genome" consisting only of adenine ($A$), thymine ($T$), and guanine ($G$). This eliminates the $C$-to-$T$ mismatch variance and allows the aligner to find the correct mapping position.
  • Post-alignment: Once the position is found, the intermediate "reduced" sequences are discarded, and the original sequences are compared to the reference to identify which original $C$ sites were conserved (methylated) or converted (unmethylated).

• Interpreting Results

  • Canonical DNA methylation occurs at CpG dinucleotides (a $C$ followed by a $G$).
  • Sites that stay as $C$ within a CpG are likely methylated.
  • Sites that stay as $C$ but are not in a CpG dinucleotide may indicate a failure in the bisulfite treatment, as non-CpG methylation is less common.

Quantitative Metrics and Statistical Confidence

• The Beta Value Calculation

  • Methylation level at a specific site is expressed as a proportion called the beta value.
  • $\text{Beta Value} = \frac{\text{Number of Methylated Reads}}{\text{Total Number of Reads}}$
  • Example: If a site has 18 methylated reads out of 20 total reads, the methylation level is $0.9$.

• Statistical Confidence and Coverage

  • Coverage refers to the total number of reads obtained at a specific genomic location.
  • Proportions alone do not tell the full story: $2/2$ methylated reads and $20/20$ methylated reads both result in a methylation level of $1.0$, but the latter provides much higher statistical confidence.
  • Higher coverage reduces the likelihood that an observation is due to a sequencing error or experimental noise.

Interpretation of Metadata Plots in Epigenomics

• Reading Metadata Graphs (Transcribed Regions)

  • Y-axis: Represents the percentage of genes with methylation in defined intervals.
  • X-axis: Often shows a "meta-graph" of transcribed regions, normalized by percentage ($0$ to $100$%) because gene lengths vary.
  • Key Finding 1 (3' Bias): Empirical data shows that methylation in expressed genes typically has a "three prime ($3'$) bias," meaning there is more methylation toward the end of the transcribed region.
  • Key Finding 2 (Pseudogenes and Non-expressed Genes): Relative to expressed genes, pseudogenes and non-expressed genes tend to have higher and more uniform methylation levels across the gene body and flanking regions.

• Entropy and Tissue Specificity

  • Entropy is a metric used to evaluate tissue specificity in gene expression.
  • A low entropy value (shifted to the left) indicates high tissue specificity.
  • A high entropy value (shifted to the right) indicates low tissue specificity.

Structural Biology of Chromatin: Nucleosomes and Histones

• Compaction vs. Accessibility

  • The human genome contains approximately $2 \text{ meters}$ of DNA per cell, which must be packed into a microscopic nucleus.
  • Structural requirement: DNA must be compacted around histones.
  • Functional requirement: Chromatin must remain accessible for transcription factors, replication proteins, and repair enzymes.

• States of Chromatin

  • Euchromatin: Less dense, more open structure. It is enriched with histone acetylation and serves as a site for active transcription.
  • Heterochromatin: Very dense, compact appearance. It contains less acetylated histones and possesses negative chromatin modifications that repress transcription.

• The Nucleosome Unit

  • The nucleosome is the fundamental unit of chromatin.
  • Structure: Approximately $147 \text{ base pairs}$ of DNA wrapped $1.75 \text{ turns}$ around a histone octamer (two copies each of core histones H2A, H2B, H3, and H4).
  • Linker Histones: H1 or H5 sit at the entry/exit points of the DNA to lock the nucleosome structure.
  • Charge Dynamics: DNA is negatively charged, while histones are positively charged, creating a strong binding affinity.

Histone Post-Translational Modifications (PTMs)

• Mechanism of Action

  • Modifications occur primarily on the histone tails.
  • These modifications alter the chemical charge of the tail, loosening or tightening the histone's grip on the DNA.
  • Common types include methylation, acetylation, and phosphorylation.

• Naming Convention

  • Example: H3K4Me1
  • H3: The name of the histone.
  • K: The single-letter code for the amino acid modified (Lysine).
  • 4: The position of the amino acid in the histone chain.
  • Me1: The type/degree of modification (Mono-methylation).

• The Histone Code

  • The concept that specific combinations of modifications act as "marks" that specify the transcriptional state of a gene.
  • Histone modifications are heritable (epigenetic) and reversible.
  • Histone Acetylation: Almost always associated with transcriptionally active genes.
  • Histone Methylation: Can be either activating or repressing depending on the site and degree of methylation.

• Recruitment of Trans Factors

  • Epigenetic marks are recognized by specific protein domains.
  • Bromo domains specifically recognize and bind to acetylated lysine residues.
  • Proteins containing these domains (e.g., transcriptional activators) are recruited to specific regions to regulate gene expression.

Technologies for Profiling Histone Modifications

• ChIP-seq (Chromatin Immunoprecipitation Sequencing)

  • The foundational technology for genome-wide profiling of histone modifications.
  • Relies on highly specific antibodies generated for every known type of modification (e.g., a specific antibody for H3K4Me1 vs. H3K4Me3).

• The ChIP-seq Workflow

  • 1. Cross-link DNA and proteins to stabilize the interaction.
  • 2. Isolate the chromatin.
  • 3. Immunoprecipitation (IP): Use specific antibodies to pull down DNA fragments associated with the targeted modification.
  • 4. Sequencing: Convert the recovered DNA into a library (Illumina or long-read) and sequence it.
  • 5. Alignment: Map reads to the reference genome to identify regions of enrichment.

• Peak Calling and Statistical Controls

  • Peaks represent genomic locations where counts of reads are significantly higher than the background.
  • Peak calling software defines the start and end of these regions using statistical models.
  • Input Controls: Used to estimate background noise in a gene-specific way.
  • Biological Replicates: Necessary to estimate variability and uncertainty in the data to distinguish biological signal from noise.

Probing Open Chromatin: ATAC-seq

• Function

  • Assay for Transposase-Accessible Chromatin using sequencing.
  • It identifies regions of open chromatin without looking for specific histone marks or DNA methylation.

• The Tn5 Mechanism

  • The assay uses the Tn5 transposase enzyme.
  • Tn5 inserts sequencing adapters specifically into regions of accessible (open) chromatin.
  • After sequencing, read "peaks" indicate active promoters, enhancers, or other regulatory regions.

Limitations of Bulk Tissue Analysis

• The Cell Type Proportion Issue

  • Most epigenomic assays (ChIP-seq, ATAC-seq, Bisulfite seq) are performed on bulk tissue samples.
  • Tissues are composed of many different cell types (e.g., the brain contains neurons, glia, immune cells).
  • A detected change in signal might not represent a change in gene regulation within cells; it may simply reflect a change in the proportion of cell types within the tissue.
  • Example: In Alzheimer's disease, a decrease in a neuronal marker might result from the loss of neurons rather than a change in the regulatory state of existing neurons.

• The Shift to Single-Cell Methods

  • To solve this, researchers now use cell sorting to isolate populations.
  • Single-cell RNA-seq, single-cell ATAC-seq, and single-nucleus methods allow the study of each individual cell's expression and chromatin state.

Epigenetics in Human Disease and Therapeutics

• Cancer Epigenetics

  • Tumors often show dysregulation in the enzymes responsible for DNA methylation and histone modification.
  • Hypermethylation: Tumor suppressor genes are frequently silenced by excessive DNA methylation in cancer types like leukemia.
  • Histone Deacetylation: Abnormal silencing of genes occurs through the activity of histone deacetylases.

• Epigenetic Therapeutics

  • Unlike genetic mutations, epigenetic changes are reversible and non-permanent.
  • Strategies involve drugs that "turn off the silence."
  • HDAC Inhibitors: Prevent histone deacetylation to keep genes in an active state.
  • DNA Methylation Inhibitors: Prevent hypermethylation to reactivate silenced tumor suppressor genes.

Questions & Discussion

• Question 1: Alignment in Bisulfite Sequencing

  • Question: Based on a simple alignment example, which of seven cytosines is methylated?
  • Context: Reference genome vs. sequencing read comparison after converting unmethylated Cs to Ts.
  • Response: The fifth position was the answer. It remained as a $C$ and was located in a CpG dinucleotide.

• Question 2: Quantitative Interpretation of Bisulfite Reads

  • Question: A CpG has $8$ reads; $6$ reads show $C$ and $2$ show $T$. Which statement is most accurate?
  • Fact: The ratio is $6/8 = 0.75$.
  • Response: The most accurate statement is that $75$% of the sequenced DNA molecules are methylated. We cannot definitively say $75$% of the cells are methylated because a bulk sample averages multiple molecules from many cells.

• Question 3: Heat Map and Expression Correlation

  • Question: In a metadata heat map showing H3K4 trimethylation (an activation mark), what does the blue color represent?
  • Context: Comparison against cluster-specific gene expression levels.
  • Response: Blue represents low H3K4 trimethylation levels. This is concluded because the clusters associated with the blue color in the heat map also show significantly lower gene expression in the corresponding box plots.

• Question 4: Bulk Tissue Dynamics

  • Question: Why is the bulk tissue signal potentially misleading?
  • Response: Bulk sampling provides an average signal. Changes observed may reflect cell type proportion changes in the sample rather than actual regulatory modifications.