kubu4/yaamini-virginica

Analysis of differentially methylated regions in Eastern oysters exposed to ambient or high pCO2 conditions.

Influence of ocean acidification on Eastern oyster (Crassostrea virginica) reproductive tissue

Background

Epigenetic modification, specifically DNA methylation, is one possible mechanism for transgenerational plasticity. Before inheritance of methylation patterns can be characterized we need a better understanding of how environmental change modifies the parental epigenome. Specifically, methylation patterns should be understood in reproductive tissue. Collaboarting with Dr. Kathleen Lotterhos' lab at Northeastern University, we examined the effect of ocean acidification on Eastern oyster (Crassostrea virginica) reproductive tissue.

The Lotterhos lab exposed twelve oysters to two different pCO₂ for four weeks at 15ºC during the summer of 2017. Half of the oysters were exposed to 400 µatm (control), and the other half to 2800 µatm. The Lotterhos lab sent gonad samples for MBDSeq to identify if different pCO₂ conditions drive differential methylation patterns. I prepared samples for bisulfite sequencing. Information about sample preparation and tangential analyses can be found in the broader project repository.

Question

Does acute exposure to elevated pCO₂ conditions induce differential methylation in Crassostrea virginica reproductive tissue?

Objectives

My goal is to identify differentially methylated regions and loci between oysters exposed to ambient and elevated pCO₂ conditions.

Methods

Overview

1. Receive sequencing data and trim files as appropriate
2. Align trimmed files to a reference bisulfite genome in bismark
3. Isolate differentially methylated loci (DML) and regions (DMR) from alignments in methylKit
4. Characterize DML and DMR with bedtools

Figure 1. Overview of methods used in this project.

Project Timeline

Week 4:

• Start bismark alignment on Mox: Protocol can be found here
• Use different samples to create working methylKit protocol. This work was done in a different Github repository.
  • Validate analysis parameters in methylKit with other samples: Results can be found here
  • Create protocol for tiling analysis in methylKit: Results can be found here

Week 5:

• Added R Markdown file for methylKit analysis
• Added Jupyter notebook for DML and DMR characterization

Week 6:

• Ran methylKit to identify DML and DMR on samples from Mox with this R Markdown script
• Characterized locations of DML and DMR in this Jupyter notebook

Week 7:

• Completed transposable elements analysis in this Jupyter notebook. Results can be found here.
• Conducted flanking analysis with bedtools flank and bedtools closest in this Jupyter notebook. Results can be found here

Week 9:

• Visualize data and complete all remaining analyses

Results

Analysis parameter validation

A full descriiption of methylKit parameter validation can be found here, with a focus on methylKit tiling analysis here.

bismark

Table 1. Mapping efficiency (%) for Bismark v.19.0 and Bowtie 2-2.3.4 (Linux x84_64 version) alignment of trimmed sample sequences to the C. virginica bisulfite genome. Mapping efficiency (%) based on different -score_min settings. For final analyses, an alignment score function of f(x) = 0 - 1.2x, where x is the read length, was set using -score_min L,0,-1.2 to set define alignment stringency and optimize mapping efficiency for all samples.

Treatment	Sample	L,0,-0.6	L,0,-0.9	L,0,-1.2
Control	1	15.5	20.2	28.8
Control	2	32.4	40.2	49.8
Control	3	37.2	45.3	53.6
Control	4	36.0	44.7	52.9
Control	5	34.6	42.9	51.7
High	6	36.7	45.0	53.8
High	7	34.6	42.9	51.4
High	8	31.7	39.0	47.6
High	9	33.0	41.2	49.9
High	10	36.6	44.9	53.0

Figure 2. Mapping efficiency (%) for Bismark v.19.0 and Bowtie 2-2.3.4 (Linux x84_64 version) alignment of trimmed sample sequences to the C. virginica bisulfite genome.

methylKit

DML:

Table 2. The mincov metric, total number of loci produced, and the number of DMLs that were at least 50% different between treatment andc control samples. More restrictive mincov metrics produced less significantly different DMLs. mincov = 3 was used in final analysis.

mincov	Total Loci	Number of Significantly Different DMLs
1	1112085	4904
3	670301	1398
5	503780	816

Figure 3. Dendogram for clustering of full sample methylation using mincov = 3 for DML.

Figure 4. Principal Components Analysis of full sample methylation using mincov = 3 for DML.

• Coverage plots for each sample generated with mincov = 3.
• BEDfile of DML identified

DMR:

Table 3. Window size, step size, total number of regions produced, and the number of DMLs that were at least 50% different between treatment and control samples. The number of regions and siginificantly different DMRs seem to be dictated by the window size, and not the step size.

Window Size (bp)	Step Size (bp)	Total Regions	Number of Significantly Different DMRs
100	100	217538	162
1000	1000	104144	118
1000	100	104144	118

Figure 5. Dendogram for clustering of full sample methylation using mincov = 3 and 100 bp window and step sizes for DMR.

Figure 6. Principal Components Analysis of full sample methylation using mincov = 3 and 100 bp window and step sizes for DMR.

• BEDfile of DMR identified

DML and DMR location characterization

A full description can be found here.

Figure 6. Full sample methylation viewed in Integrative Genome Viewer.

DML:

• DML-exon overlaps
• DML-intron ovleraps
• DML-mRNA overlaps
• Unique genes in DML-mRNA overlaps
• DML-TE overlaps

Table 4. Location of differentially methylated loci (DML) in various genomic features from BEDtools intersect v2.26.0. Genome feature files were downloaded from NCBI. The C. virginica genome has 60,201 genes total. For each locus, hypermethylated refers to significantly higher methylation in treatment samples, while hypomethylated indicates significantly lower methylation. Transposable elements refers to those identified using C. gigas as the species designation.

Genomic Feature	Result
Total DML	1398
Hypomethylated DML	747
Hypermethylated DML	651
Total genes with DML	2,683
DML in mRNA coding regions	1263 (90.34%)
DML in exons	786 (56.22%)
DML in introns	498 (35.62%)
DML in transposable elements	91 (6.51%)

DMR:

• DMR-exon overlaps
• DMR-intron ovleraps
• DMR-mRNA overlaps
• Unique genes in DMR-mRNA overlaps
• DMR-TE overlaps

Table 5. Location of differentially methylated regions (DMR) in various genomic features from BEDtools intersect v2.26.0. Regions were identified in a tiling window analysis in methylKit v.1.7.9 in R Genome feature files were downloaded from NCBI. The C. virginica genome has 60,201 genes total. For each 100 bp region, hypermethylated refers to significantly higher methylation in treatment samples, while hypomethylated indicates significantly lower methylation. Transposable elements refers to those identified using C. gigas as the species designation.

Genomic Feature	Result
Total DMR	162
Hypomethylated DMR	23
Hypermethylated DMR	139
Total genes with DMR	305
DMR in mRNA coding regions	139 (85.80%)
DMR in exons	64 (39.51%)
DMR in introns	112 (69.14%)
DMR in transposable elements	23 (14.20%)

Transposable elements:

Table 6. Percent overlap between transposable element and exons, introns, and mRNA coding regions. Transposable elements refers to those identified using C. gigas as the species designation.

Genomic Feature	Overlap
mRNA Coding Regions	7.79
Exons	6.00
Introns	14.2

Flanking analysis

A full description can be found here.

bedtools flank

1000 bp flanks for mRNA coding regions

DML:

• DML-upstream flank overlaps
• DML-downstream flank overlaps

DMR:

• DMR-upstream flank overlaps
• DMR-downstream flank overlaps

bedtools closest

DML:

Closest non-overlapping DML to mRNA coding regions

DMR:

Closest non-overlapping DMR to mRNA coding regions

Next Steps

1. Determine if a formal gene enrichment is necessary
2. If necessary, select the most appropriate gene enrichment method
3. Describe functions of most interesting genes with DML and DMR
4. Update draft manuscript

Repository Structure

analyses

R code and output from multiple analyses. Each analysis will be in its own subdirectory.

• 2018-10-25-MethylKit: R Markdown file and output from methylKit identification of DML and DMR.
• 2018-11-01-DML-and-DMR-Analysis: Output from bedtools characterization of DML and DMR location in the C. virginica genome

data

Raw data used for project analyses, as well as links to data files.

images

Images generated outside of standard analyses.

notebooks

Jupyter notebooks that detail reproducible methods for data analysis.

• 2018-11-01-DML-and-DMR-Analysis.ipynb: Pipeline for bedtools analysis of DML and DMR locations in various genome feature tracks. Includes option to specify variable path names for easy reproducibility.

scripts

Scripts used for Mox.

• 2018-10-31-Revised-Bismark-Parameters-Samtools: Used to run bismark alignment on MBD-Seq data. This is an edited version of 2018-10-12-Revised-Bismark-Parameters that does not redirect standard error to a new file and includes an explicit path to bowtie2 and samtools in the alignment step.

tutorials

Data, notebooks, and analyses from class tutorials.

• BASH Tutorial in Jupyter Notebook
• BASH Tutorial in R Markdown