We will use a small subset of the original data. If we would do the computation on the orginal data the computation time for a tutorial is too long. To show you all necessary steps for Methyl-Seq we decided to use a subset of the data set. In a second step we use precomputed data from the study to show you different levels of methylation. We will consider samples from normal breast cells (NB), fibroadenoma (noncancerous breast tumor, BT089), two invasive ductal carcinomas (BT126, BT198) and a breast adenocarcinoma cell line (MCF7).

This tutorial is based off of Lin et al. 2015. The data we use in this tutorial is available at Zenodo.

Agenda

In this tutorial, we will deal with:

1. Data upload
2. Quality Control
3. Alignment
4. Methylation bias and metric extraction
5. Visualization
6. Metilene

Data upload

We will start by loading the example dataset which will be used for the tutorial into Galaxy

Hands-on: Get the data into Galaxy

1. Create a new history

   Tip: Creating a new history

   To create a new history simply click the new-history icon at the top of the history panel:

   

2. Import the two example datasets from Zenodo or the shared data library:

   https://zenodo.org/record/557099/files/subset_1.fastq
   https://zenodo.org/record/557099/files/subset_2.fastq
   

   Tip: Importing via links

   • Copy the link location

   • Click galaxy-upload Upload Data at the top of the tool panel

   • Select galaxy-wf-edit Paste/Fetch Data

   • Paste the link(s) into the text field

   • Press Start

   • Close the window

   Tip: Importing data from a data library

   As an alternative to uploading the data from a URL or your computer, the files may also have been made available from a shared data library:

   1. Go into Data (top panel) then Data libraries
   2. Navigate to the correct folder as indicated by your instructor.
      • On most Galaxies tutorial data will be provided in a folder named GTN - Material –> Topic Name -> Tutorial Name.
   3. Select the desired files
   4. Click on Add to History galaxy-dropdown near the top and select as Datasets from the dropdown menu

   5. In the pop-up window, choose

      • “Select history”: the history you want to import the data to (or create a new one)
   6. Click on Import

Quality Control

The first step in any analysis should always be quality control. We will use the FastQC tool to asses the quality of our reads and determine if we need to perform any data cleaning before proceeding with our analysis.

Hands-on: Quality Control

1. FastQC tool with the following parameters:
   • param-files “Raw read data from your current history”: subset_1.fastq.gz and subset_2.fastq.gz

   Tip: Select multiple datasets

   1. Click on param-files Multiple datasets
   2. Select several files by keeping the Ctrl (or COMMAND) key pressed and clicking on the files of interest

2. Go to the web page result page and have a closer look at ‘Per base sequence content’

   

   Question

   1. Note the GC distribution and percentage of “T” and “C”. Why is this so weird?
   2. Is everything as expected?

   Solution

   1. The attentive audience of the theory part knows: Every C-meth stays a C and every normal C becomes a T during the bisulfite conversion.
   2. Yes it is. Always be careful and have the specific characteristics of your data in mind during the interpretation of FastQC results.

Alignment

Hands-on: Mapping with bwameth

We will map now the imported dataset against a reference genome.

1. bwameth tool with the following parameters:
   • Select for the option Select a genome reference from your history or a built-in index? Use a built-in index and here the human hg38 genome.
   • Choose for the option Is this library mate-paired? Paired-end and use the two imported datasets as an input. Compute now the alignment. Please notice that depending on your system this computation can take some time. If you want to skip this, we provide for you a precomputed alignment. Import aligned_subset.bam to your history.

   Question

   Why we need other alignment tools for bisulfite sequencing data?

   Solution

   You may have noticed that all the C’s are C-meth’s and a T can be a T or a C. A mapper for methylation data needs to find out what is what.

Methylation bias and metric extraction

Hands-on: Methylation bias

In this step we will have a look at the distribution of the methylation and will look at a possible bias.

1. MethylDackel tool with the following parameters:
   • Choose at the first option Load reference genome from Local cache and for Using reference genome the value hg38.
   • Select for the option sorted_alignments.bam the computed bam file of step 4 of the bwameth alignment.
   • Use for What do you want to do? the value Determine the position-dependent methylation bias in the dataset, producing diagnostic SVG images.
   • Set the parameters By default, if only one read in a pair aligns (a singleton) then it's ignored. and By default, paired-end alignments with the properly-paired bit unset in the FLAG field are ignored. Note that the definition of concordant and discordant is based on your aligner settings. to Yes.

   

   Question

   1. Consider the original top strand output. Is there a methylation bias?
   2. If we would trim, what would be the start and the end positions?

   Solution

   1. The distribution of the methylation is more or less equal. Only at the start and the end we could trim a bit but a +- 5% variation is acceptable.
   2. To trim the reads we would include for the first strand only the positions 0 to 145, for the second 6 to 149.

Hands-on: Methylation extraction with MethylDackel

We will extract the methylation on the resulting BAM file of the alignment step. We need this to create a methylation level plot in the next step.

1. MethylDackel tool with the following parameters:
   • Choose at the first option Load reference genome from the value: Local cache and for Using reference genome the value: hg38.
   • Select for the option sorted_alignments.bam the computed bam file of step 4 of the bwameth alignment.
   • Use for What do you want to do? the value Extract methylation metrics from an alignment file in BAM/CRAN format.
   • Choose Yes for the option Merge per-Cytosine metrics from CpG and CHG contexts into per-CPG or per-CHG metrics.
   • Set the parameter Extract fractional methylation (only) at each position. This is mutually exclusive with --counts, --logit, and --methylKit to Yes.
   • All other options use the default value.

Visualization

Hands-on

In this step we want to visualize the methylation level around all TSS of our data. When located at gene promoters, DNA methylation is usually a repressive mark.

1. Wig/BedGraph-to-bigWig tool with the following parameters:

   • Use the result of MethylDackel to transform it to a bigWig file.

     Tip: Database edit

     It can happen that you can not select the correct input file. In this case you have to add meta information about the used genome to the file.

     • Click on the pencil of the correct history item.
     • Change Database/Build: to the genome you used.
     • In our case the correct genome is Human Dec. 2013 (GRCh38/hg38) (hg38).

2. computeMatrix tool with the following parameters:
   • Use the file CpGIslands.bedas Regions to plot and the in the previous step created bigwig file as the score file.
   • Use for the option computeMatrix has two main output options the value reference-point.
3. plotProfile tool with the following parameters:
   • Choose for Matrix file from the computeMatrix tool the computed matrix from the tool computeMatrix.

The output should look like this:

Lets see how the methylation looks for a few provided files:

1. Galaxy tool: Import the files NB1_CpG.meth.bedGraph from the data library
2. Wig/BedGraph-to-bigWig tool with the following parameters:
   • Use the imported file to transform it to a bigWig file.

   Question

   The execution fails. Do you have an idea why?

   Solution

   A conversion to bigWig would fail right now, probably with some error message like hashMustFindVal: '1' not found. The reason is the source of the reference genome which was used. There is ensembl and UCSC as sources which differ in naming the chromosomes. Ensembl is using just numbers e.g. 1 for chromosome one. UCSC is using chr1 for the same. Be careful with this especially if you have data from different sources. We need to convert this.

   Comment: UCSC - Ensembl convert

   • Download the Replace information file for hg38 chromosome: Download and import it to Galaxy.
   • Replace column tool:
     • Choose for File in which you want to replace some values the previous used NB1_CpG.meth.bedGraph file and for Replace information file conversion file. For Which column should be replaced? choose Column: 1, for Skip this many starting lines a 1 and for Delimited by Tab.

3. To save compute time we prepared the converted files for you. Import the files: NB1_CpG.meth_ucsc.bedGraph, NB2_CpG.meth_ucsc.bedGraph, BT089_CpG.meth_ucsc.bedGraph, BT126_CpG.meth_ucsc.bedGraph, BT198_CpG.meth_ucsc.bedGraph and MCF7_CpG.meth_ucsc.bedgraph.
4. Compute the matrix and plot the profile as described above.

More information about deepTools can be found here: https://deeptools.readthedocs.io

Metilene

Hands-on: Metilene

With metilene it is possible to detect differentially methylated regions (DMRs) which is a necessary prerequisite for characterizing different epigenetic states.

1. Galaxy tool: Import from the data library the files NB1_CpG.meth.bedGraph, NB2_CpG.meth.bedGraph and BT198_CpG.meth.bedGraph.
2. Metilene tool:
   • Choose for the first option Input group 1 the imported files starting with NB and for Input group 2 the imported files BT198_CpG.meth.bedGraph.
   • Select for the option BED file containing regions of interest the imported BED file CpGIslands.bed.

3. More information about metilene can be found here: https://www.bioinf.uni-leipzig.de/Software/metilene

   Question

   Have a look at the produced pdf document. What is the data showing?

   Solution

   It shows the distribution of DMR differences, DMR length in nucleotides and number CpGs, DMR differences vs. q-values, mean methylation group 1 vs. mean methylation group 2 and DMR length in nucleotides vs. length in CpGs

You've Finished the Tutorial

Please also consider filling out the Feedback Form as well!

Key points

• The output of a methylation NGS is having a different distribution of the four bases. This is caused by the bisulfite treatment of the DNA.

• If there is a different level of methylation in the loci of a gene this can be a hint that something is wrong.

• To get useful results you need – data, data and data!

Frequently Asked Questions

References

1. Lin, I.-H., D.-T. Chen, Y.-F. Chang, Y.-L. Lee, C.-H. Su et al., 2015 Hierarchical Clustering of Breast Cancer Methylomes Revealed Differentially Methylated and Expressed Breast Cancer Genes (O. El-Maarri, Ed.). PLOS ONE 10: e0118453. 10.1371/journal.pone.0118453

Feedback

Did you use this material as an instructor? Feel free to give us feedback on how it went.
Did you use this material as a learner or student? Click the form below to leave feedback.

Citing this Tutorial

1. Joachim Wolff, Devon Ryan, DNA Methylation data analysis (Galaxy Training Materials). https://training.galaxyproject.org/training-material/topics/epigenetics/tutorials/methylation-seq/tutorial.html Online; accessed TODAY
2. Hiltemann, Saskia, Rasche, Helena et al., 2023 Galaxy Training: A Powerful Framework for Teaching! PLOS Computational Biology 10.1371/journal.pcbi.1010752
3. Batut et al., 2018 Community-Driven Data Analysis Training for Biology Cell Systems 10.1016/j.cels.2018.05.012

BibTeX

@misc{epigenetics-methylation-seq,
author = "Joachim Wolff and Devon Ryan",
	title = "DNA Methylation data analysis (Galaxy Training Materials)",
	year = "",
	month = "",
	day = "",
	url = "\url{https://training.galaxyproject.org/training-material/topics/epigenetics/tutorials/methylation-seq/tutorial.html}",
	note = "[Online; accessed TODAY]"
}
@article{Hiltemann_2023,
	doi = {10.1371/journal.pcbi.1010752},
	url = {https://doi.org/10.1371%2Fjournal.pcbi.1010752},
	year = 2023,
	month = {jan},
	publisher = {Public Library of Science ({PLoS})},
	volume = {19},
	number = {1},
	pages = {e1010752},
	author = {Saskia Hiltemann and Helena Rasche and Simon Gladman and Hans-Rudolf Hotz and Delphine Larivi{\`{e}}re and Daniel Blankenberg and Pratik D. Jagtap and Thomas Wollmann and Anthony Bretaudeau and Nadia Gou{\'{e}} and Timothy J. Griffin and Coline Royaux and Yvan Le Bras and Subina Mehta and Anna Syme and Frederik Coppens and Bert Droesbeke and Nicola Soranzo and Wendi Bacon and Fotis Psomopoulos and Crist{\'{o}}bal Gallardo-Alba and John Davis and Melanie Christine Föll and Matthias Fahrner and Maria A. Doyle and Beatriz Serrano-Solano and Anne Claire Fouilloux and Peter van Heusden and Wolfgang Maier and Dave Clements and Florian Heyl and Björn Grüning and B{\'{e}}r{\'{e}}nice Batut and},
	editor = {Francis Ouellette},
	title = {Galaxy Training: A powerful framework for teaching!},
	journal = {PLoS Comput Biol}
}

                   

Galaxy Administrators: Install the missing tools

You can use Ephemeris's shed-tools install command to install the tools used in this tutorial.

shed-tools install [-g GALAXY] [-a API_KEY] -t <(curl https://training.galaxyproject.org/training-material/api/topics/epigenetics/tutorials/methylation-seq/tutorial.json | jq .admin_install_yaml -r)

Alternatively you can copy and paste the following YAML

---
install_tool_dependencies: true
install_repository_dependencies: true
install_resolver_dependencies: true
tools:
- name: deeptools_compute_matrix
  owner: bgruening
  revisions: fd1275e01605
  tool_panel_section_label: deepTools
  tool_shed_url: https://toolshed.g2.bx.psu.edu/
- name: deeptools_plot_profile
  owner: bgruening
  revisions: aac8444d6681
  tool_panel_section_label: deepTools
  tool_shed_url: https://toolshed.g2.bx.psu.edu/
- name: pileometh
  owner: bgruening
  revisions: 906db57d5d65
  tool_panel_section_label: Peak Calling
  tool_shed_url: https://toolshed.g2.bx.psu.edu/
- name: replace_column_by_key_value_file
  owner: bgruening
  revisions: cc18bac5afdb
  tool_panel_section_label: Text Manipulation
  tool_shed_url: https://toolshed.g2.bx.psu.edu/
- name: text_processing
  owner: bgruening
  revisions: e39fceb6ab85
  tool_panel_section_label: Text Manipulation
  tool_shed_url: https://toolshed.g2.bx.psu.edu/
- name: text_processing
  owner: bgruening
  revisions: d698c222f354
  tool_panel_section_label: Text Manipulation
  tool_shed_url: https://toolshed.g2.bx.psu.edu/
- name: fastqc
  owner: devteam
  revisions: 484e86282f4b
  tool_panel_section_label: FASTA/FASTQ
  tool_shed_url: https://toolshed.g2.bx.psu.edu/
- name: bwameth
  owner: iuc
  revisions: a6ea26c1f225
  tool_panel_section_label: Mapping
  tool_shed_url: https://toolshed.g2.bx.psu.edu/
- name: metilene
  owner: rnateam
  revisions: 37abd09a0ae6
  tool_panel_section_label: Epigenetics
  tool_shed_url: https://toolshed.g2.bx.psu.edu/

Feedback

5 stars	7
4 stars	2
3 stars	1
1 stars	3

October 2024

• 3 stars: Liked: option along with version numbres help
• 4 stars: Disliked: more example

April 2024

• 4 stars: Disliked: Actually, I encountered an error while transforming the MethylDackel file to a bedgraph file. The error message was: "unrecognized line 1 of trackless: chr1 5190715 5190717 100 1 0". After that, I wasn’t able to replicate the subsequent steps. I suggest providing clearer instructions for each step. Additionally, I would like to know more about how to group the data for analysis. This part feels a bit too concise. Overall, this material has been helpful for me to apply it to my own data, although I may need to explore additional tools&Analysis on my own. Thank you! And I was wondering if it is possible to build a comprehensive analysis workflow for Methyl-BS-Seq data in Galaxy similar to the cross-package Bioconductor workflow for analyzing methylation array data (published in F1000Research). :)

June 2022

• 1 stars: Liked: nothing Disliked: everything

January 2022

• 1 stars: Disliked: Data set can not be found

November 2021

• 1 stars: Liked: The programs were not available at Galaxy

September 2021

• 5 stars: Disliked: It would be great if this can be updated based on the newest version of the packages used in the tutorial. Some of the parameters cannot be set as shown in this tutorial or should be set in a different way.
• 5 stars: Disliked: Instructions about finding tutorial data in a Data Library need to be updated. In this tutorial and probably others. Tutorial data is now nested at usegalaxy.* servers -- and that change is confusing some learners. I've been telling people to search data libraries with the keyword "GTN" then to navigate down through the topic to the specific tutorial. Not all public servers will host the data in Data Libs, so even that advice needs a tune up to be consistent/accurate across tutorials. (@jennaj)

March 2020

• 5 stars: Liked: The degree of detail. I loved that even the parameters for the different tools were explained. I don't know if this is comparable to a real experiment analysis, but it surely felt as that. Congrats! Disliked: The Genrich tool is only available at the GalaxyEurope server. It would be nice that a warning would be given at the beginning of the tutorial, so that the whole analysis could be done there and avoid migrating data between Galaxy servers.

February 2020

• 5 stars: Liked: The tutorial is self explanatory and very easy to follow for individual hands on

September 2019

• 5 stars: Liked: The good explanations during hands-on training Disliked: Maybe too different analysis for one day (HiC and epigentics), maybe a longer session for each would help understanding