DMRichR v1.7.1

Minor updates and trigger for Zenodo

DMRichR v1.7.0

New functions:

• DMRichR::DM.R() has been parsed from the DM.R executable. The entire DMRichR pipeline can now be run as a single function directly in R. The option to use the DM.R executable through command line is still supported.
• DMRichR::prepareHOMER() and DMRichR::HOMER() allow you to test for the enrichment of transcription factor motifs within your DMRs using HOMER. The main function will only run if HOMER is detected in your path and your genome of interest is already installed.

Improvements:

• DMRichR now has a pkgdown website that offers improved documentation: https://ben-laufer.github.io/DMRichR/
• GitHub actions to perform R CMD check on ubuntu, macOS, and windows.
• DMRichR::annotateRegions() now supports annotating coordinates from analyses outside of DMRichR.

Bug fixes:

• DMRichR::annotationDatabases() now assigns the EnsDb to the proper environment so that other functions can use it.

DMRichR v1.6.1

New Features:

• DMRichR::slimGO() utilizes rrvgo and replaces DMRichR::REVIGO() due to REVIGO website update
• Support for ensembldb (--ensembl argument for DM.R), which improves annotations for non-model organisms
• DMRichR::getCpGs() is a new CpG island/feature annotation system that supports almost all organisms
• Plotting of exons and CpG feature tracks in the DMR and block plots for almost all organisms via DMRichR::getCpGs() and DMRichR::getExons()
• Interactive Global, 20Kb Window, and CpG Island MDS plots from Glimma

Improvements:

• The globalPlots approach has been redone to create separate functions to extract (DMRichR::windows(), DMRichR::CpGs(), and DMRichR::CGi()) and plot (DMRichR::PCA() and DMRichR::densityPlot())
• DMRichR::processBismark() will now automatically convert ranges to UCSC style, drop sex chromosomes if sexCheck == TRUE and both sexes are detected, and assign colors for DMR and block plots
• DMRichR::imprintOverlap() fix for non-humans
• DMRichR::GOfuncR() now supports analyses outside of DM.R that do not have exact overlap between the DMRs and background regions
• An argument to not run GOfuncR (--GOfuncR) has been added to DM.R
• enrichR now selects organism specific sites/databases if available
• rGREAT has been changed to "oneClosest" rule
• Gene ontologies will only run for all DMRs, the additional stratification of hypermethylated and hypomethylated has been removed
• DMRichR::GOplot() now has increased font size and text truncation for long terms

DMRichR v1.5.0

New Features:

• The DMRichment family of functions has been added and serves as an homage to the one of the original goals of DMRichR. DMRichCpG() and DMRichGenic() test for signficant enrichments of the DMRs (relative to background regions) within CpG and gene region (genic) annotations via a Fisher's exact test with FDR correction. DMRichPlot() plots the results of both types of enrichment tests and DMparseR() can be used to create a DMRichPlot() with facets for all DMRs, hypermethylated DMRs, and hypomethylated DMRs. Finally, imprintOverlap() tests for enrichments within human imprinted genes.
• singleCpGPCA() has been introduced.
• Added the maxBlockPerms option to DM.R. Default is 10.

Improvements:

• Added CpG island statistics to globalStats() for hg38, hg19, mm10, mm9, rn6.
• REVIGO() has been parsed from GOplot() so that the full list of slimmed gene ontologies can be saved. This also enables use with other pipelines that utilize GREAT, GOfuncR or enrichr.
• Added average length and average number of CpGs to the subtitle of DMReport().
• Houseman() has been parsed from DM.R and the experimental cell composition estimation pipeline has been tidied.
• tryCatch() is now used for plotDMRs2() to prevent a plotting error from stopping the program .
• Messaging, error checks, and documentation have been updated.

DMRichR v1.4.0

• Added the sexCheck argument to processBismark() to confirm the sex of samples by using k-means clustering of the ratio of coverage for the Y chromosome and X chromosome. Developed by @hyeyeon-hwang.
• GOplot() will now also slim GO terms using REVIGO and rank the terms based on dispensability.
• GOfuncR() now supports all genomes.
• methylLearn() heatmaps now display Z-scores and gene region annotations. Developed by @hyeyeon-hwang.
• Improvements to densityPlot() and PCA().
• Improvements to imports to reduce install time and remove install warnings.
• Moved Bioconductor annotation database loading from DM.R to DMRichR::annotationDatabases().
• Parallelized GREAT and GOfuncR and added enrichr:::.onAttach() to DM.R.
• Simplification of RData from DM.R.
• Major documentation updates that give an overview of the DM.R workflow and key DMRichR functions.

DMRichR v1.3.0

Key updates:

1. Expanded supported genomes to: hg38, hg19, mm10, mm9, rheMac10, rheMac8, rn6, danRer11, galGal6, bosTau9, panTro6, dm6, canFam3, susScr11, and TAIR9.

2. Resource optimization: DM.R should use between 10-20 cores and 64-256 GB of RAM.

3. Optimized block calling parameters.

4. densityPlot() is now for single CpGs instead of 20 KB windows, which makes it more similar to array output.

5. Added a beta feature for cell composition estimation in human whole blood using two separate methods: Houseman and methylCC. CCstats() and CCplot() are novel functions for downstream statistical analysis and visualization. The feature can be utilized by adding --cellComposition = TRUE to the DM.R call.

6. Tidying of the annotation pipeline to add gene symbols, percent difference, and directionality to regions of interest, where annotateRegions() was developed from previous helper functions and code in DM.R

7. Added annotations for blocks, where DMReport() now supports blocks and DM.R generates an HTML block report

8. Gene ontology analyses are now performed for all DMRs, hypermethylated DMRs, and hypomethylated DMRs.

9. Added arrayLift() to liftOver Illumina array (EPIC, 450K, and 27K) CpG IDs to hg38 coordinates.

10. Added LOLA enrichment testing and heatmaps for the ChromHMM 15-state model and Roadmap Epigenomics core histone modifications for hg38. Currently, this will only run on the UC Davis Cluster due to needing large external databases; however, an advanced user can download the databases and make minor modifications to the functions [chromHMM(), chromHMM_heatmap(), roadmap(), and roadmap_heatmap()] to refer to their local copy.

11. Tidying, minor bug fixes, and documentation updates.

DMRichR v1.2.0

New functions and critical maintenance updates have been made alongside many minor improvements.

The two new key functions are:

1. methylLearn() was developed by @hyeyeon-hwang and utilizes two machine learning algorithms to perform feature selection and identify the top common DMRs. It produces an html report and a heatmap.

2. smoothPheatmap()replaces smoothHeatmap(), where it plots traditional Z-scores and also offers improved aesthetics for multiple covariate annotations.

The two key updates are:

1. processBismark() has been substantially improved by @cemordaunt and the perGroup argument now considers covariate combinations when choosing the percent of samples per a group for CpG coverage filtering.

2. The gene ontology pipes have also been updated to support the latest versions of GREAT and Enrichr.

DMRichR v1.1.0

Many of the existing functions from the DM.R executable have been parsed, tidied, and documented to create the DMRichR package. A number of new functions have been added as well. DM.R still serves as the main executable, which calls these functions; however, the R data can be loaded later on and the existing functions can have their arguments modified to customize the output.

DMRichR v1.0

This a stable version that is similar to what was used in the 2019 Epigenetics manuscript.