mint v0.2.0

• Simplify project setup by adding parameters to init.R.
• Split methylSig analysis by chromosome to deal with memory issues for large WGBS runs (we will return to this solution in a future release).
• Implement simple classification for pulldown_compare.
• Refactor annotatr related code. Currently works with annotatr v0.99.5 and annotatr.data v0.99.4.
• Turn off random regions for cases where it is too computationally cumbersome.
• Remove *_align rules as prereqs for *_compare rules.
• Add intron/exon boundaries to default annotatr analyses.
• Changed default bowtie2 run parameters to include --no-unal.

mint v0.1.5

CpGs randomized to CpGs

• In the first implementation of region randomization CpGs were not required to be randomized to CpGs. Now we use the CpG report from bismark_methylation_extractor as the universe of CpGs and use regioneR::resampleRegions with this universe instead of annotatr::randomize_regions (a wrapper for regioneR::randomizeRegions).
• Committed in 534bd70.

mint v0.1.4

Improve UCSC Genome Browser tracks

• In previous versions, the classification tracks could have multiple adjacent regions with the same classification, making for a mess, as below.

• Now use bedtools merge within each classification type to avoid this. The result is as below.

mint v0.1.3

Bugfixes

• Remove hard-coded hg19 in randomize_regions()
• Fixed a typo for simple classification category hypo_mc.
• Use http instead of https for Bioconductor installs.

mint v0.1.2

Dependencies

• PePr is now at v1.1.10.
• Updated VERSIONS.md to be in alphabetical order.

makefile and config.mk changes

• Improved makefile and config.mk organization and documentation.
  • More granular rules to do QC before committing to alignments:
    • make bisulfite_raw_fastqc
    • make bisulfite_trim
    • make bisulfite_trim_fastqc
    • make pulldown_raw_fastqc
    • make pulldown_trim
    • make pulldown_trim_fastqc
• Create project specific tmp folder for sorting, etc.
• Remove PBS script creation.

Parameters

• Fix --mfold parameter formatting for macs2.
• Autofill for --gsize parameter for macs2 when genome is hg19, hg38, mm9, or mm10.
• Add parameters for group names in all comparisons that propagate to classifications and annotatr output. For example, in the figure below the two groups are IDH2mut and NBM rather than chip1 and chip2 as before.

New modules

• Added multiqc module that creates output in summary/reports/multiqc and splits the reports between bisulfite and pulldown samples.
  • make bisulfite_multiqc
  • make pulldown_multiqc

New plots

• Add fold change plots to annotatr output for PePr and macs2 peaks. See below for examples.

mint v0.1.1

This is the first official release of mint. The release notes are excerpts from this version's README.

Summary

The mint pipeline analyzes single-end reads coming from sequencing assays measuring DNA methylation. The pipeline analyzes reads from both bisulfite-conversion assays such as WGBS and RRBS, and from pulldown assays such as MeDIP-seq, hMeDIP-seq, and hMeSeal. Moreover, with data measuring both 5-methylcytosine (5mc) and 5-hydroxymethylcytosine (5hmc), the mint pipeline integrates the two data types to classify genomic regions of 5mc, 5hmc, a mixture, or neither.

The mint pipeline is executed with make and includes configurable steps for:

• Quality control (FastQC)
• Adapter and quality trimming (trim_galore)
• Alignment (bismark and bowtie2)
• Sample-wise quantification (bismark and macs2)
• Group-wise differential methylation detection (methylSig and PePr)
• Classification
  • of samples into regions of no, low, medium, or high methylation
  • of 5mc + 5hmc sample-wise integration into regions of 5mc, 5hmc, a mixture, or neither
  • of group comparisons into regions of hyper/hypo DMR or hyper/hypo DhMR
  • of 5mc + 5hmc group-wise integration into regions of hyper/hypo DMR, hyper/hypo DhMR, a mixture, or neither
• Genomic annotation and visualization of methylation quantifications and classifications (annotatr)
• Visualization in the UCSC Genome Browser

Supported Experiments and Designs

The mint pipeline supports either:

• a hybrid setup, with data from bisulfite-conversion experiments representing 5mc + 5hmc methylation (RRBS, WGBS, etc.) and from pulldown experiments representing 5hmc methylation (hMeDIP-seq, hMeSeal, etc.), or
• a pulldown setup, with data representing 5mc methylation (MeDIP-seq, etc.) and 5hmc methylation (hMeDIP-seq, hMeSeal, etc.).

And the mint pipeline can analyze:

• sample-wise, where no groups are present for comparison, and 5mc / 5hmc integration is done per-sample, and/or
• comparison-wise, where two or more groups are tested for differential methylation, and 5mc / 5hmc differential methylation integration is done per-group.

Dependencies

The mint pipeline is dependent on several software packages to carry out its analysis, integration, annotation, and visualization. Links to appropriate versions are included in the list below.

• bedtools v2.25.0
• bedops v2.4.14
• bismark v0.16.1
• bowtie2 v2.2.4
• macs2 v2.1.0.20140616
• PePr v1.1.5
• R >= v3.2.5
  • annotatr v0.7.3
  • methylSig v0.4.3
• samtools v0.1.19
• trim_galore v0.4.1
• cutadapt v1.9.1
• FastQC v0.11.5