devgru

A developmental R environment for a suite of GenomicRanges utilities

devgru provides tools for cancer genomics analysis within the GenomicRanges universe: format-aware file readers, conversions between data.table, GRanges, GRangesList and GInteractions, seqinfo harmonization, chromosome-parallel workflows, copy number, structural variant, and RNA-seq analysis helpers, QC diagnostics, and plotting. By default everything is anchored to the autosomes and sex chromosomes of hg38 (see gUtils::hg_seqlengths()).

Full documentation and the function reference are available at https://pblaney.github.io/devgru/.

Installation

You can install the current development version of devgru with:

devtools::install_github("pblaney/devgru")

Several dependencies are only available on GitHub (mskilab-org/gUtils, mskilab-org/fishHook, caravagnalab/easypar, caravagn/pio, showteeth/ggpie, and pblaney/gGnome); these are declared as Remotes and resolved automatically by devtools/remotes. Packages listed under Suggests (e.g. DESeq2, fgsea, DNAcopy, InteractionSet) are only needed for the specific workflows that use them.

Suite of Packages

The current suite of utilities for devgru can be quickly installed, if needed, and loaded with:

library(devgru)
kit_loadout()

# Install/upgrade any missing or outdated packages in the kit first
kit_loadout(update_kit = TRUE)

GenomicRanges Core

• BSgenome.Hsapiens.UCSC.hg38
• GenomicRanges
• GenomeInfoDb
• data.table
• mskilab-org/gUtils
• VariantAnnotation
• rtracklayer
• Biostrings
• S4Vectors

Utility Core

• dplyr
• stringr
• readr
• ggplot2
• ggsci
• paletteer
• scico
• flextable
• mclust
• parallel
• doParallel
• foreach
• R.utils

Function Overview

Category	Functions
Environment setup	kit_loadout() loads the package kit; chrompar() runs a function in parallel by chromosome
Readers	read_bed_file(), read_bedpe_file(), read_gtf_file(), read_maf_file(), read_vcf_file() (output as GRanges or data.table), plus aggregate_these() to detect the format of and combine many files into one data.table
Conversions	dt_to_gr(), dt_to_gi(), gi_to_dt(), gi_to_grl(), grl_to_gi(), gr_to_seg()
GenomicRanges utilities	gr_refactor_seqs(), grl_refactor_seqs(), gi_refactor_seqs() to harmonize seqinfo; gr_flank() for boundary-aware flanking; gr_sanitycheck() and dt_sanitycheck() for input validation
Copy number	get_fragcounter_segmentation(), get_cbs_per_chromosome(), get_cov_gap_imputation(), get_imputed_gaps_per_chromosome(), get_corrected_cnv_profile()
Mutations & structural variants	get_vaf(), get_maf_lite(), get_maf_nonsynonymous(), get_igh_trx_subtype() (canonical IGH translocations in multiple myeloma)
RNA-seq	get_deseq2_diff_expr(), get_fgsea_pathway_enrichment(), get_protein_coding_genes()
QC diagnostics	get_qc_diagnostics_alignment() (Alfred), get_qc_diagnostics_snvindel(), get_jabba_qc_diagnostic()
Plotting	geom_deseq2_volcano(), geom_gsea_enrichment(), geom_gap_imputation(), geom_jabba_stats(), copynumber_palettier()
CLI helpers	cli_stopwatch_start(), cli_stopwatch_end(), function_cli_intro(), cli_pio_colorscheme()

See the reference page for full details on each function.

Quick Start

Read genomic files straight into harmonized GRanges (or a data.table with output_type = "dt"):

library(devgru)

peaks_gr <- read_bed_file("peaks.bed.gz")
vcf_gr <- read_vcf_file("sample.vcf.gz", cpus = 4)
sv_dt <- read_bedpe_file("sample.bedpe")

# Collect and combine all MAFs in a directory into a single data.table
cohort_maf <- aggregate_these(path_to_files = "mafs/",
                              pattern_to_grab = "\\.maf$",
                              add_uniq_id = TRUE)

Harmonize seqinfo across objects, e.g. after lifting over from hg19 to hg38:

GenomeInfoDb::seqinfo(example_dnase_hg38)
GenomeInfoDb::seqinfo(gr_refactor_seqs(example_dnase_hg38))

Add flanks to a region without crossing a boundary such as the centromere:

notch2 <- gUtils::parse.gr("chr1:119911553-120069662")
chr1_p_arm <- gUtils::`%Q%`(chromosome_arms_hg38, seqnames == "chr1" & arm == "p")

gr_flank(input_gr = notch2,
         start_flank = 3e6,
         end_flank = 3e6,
         start_flank_boundary = GenomicRanges::start(chr1_p_arm),
         end_flank_boundary = GenomicRanges::end(chr1_p_arm))

Run larger workflows, such as CBS segmentation of a fragCounter read depth profile parallelized by chromosome:

fragcounter_seg <- get_fragcounter_segmentation(
    path_to_fragcounter_profile = "sample.fragcounter.rds",
    cpus = 4)

Bundled Data

devgru ships with hg38 reference annotations used internally by its workflows, as well as small demo datasets for examples:

• Reference genomic feature annotations: chromosome_arms_hg38, exclusion_regions_hg38, gene_body_hg38 (Ensembl v108),
• Myeloma-specific genomic features: myeloma_translocation_ig_loci_hg38, myeloma_translocation_partner_loci_hg38
• Demo data: DNase I hypersensitivity sites (example_dnase_hg38, braf_dnase_demo_dt_hg19/hg38, kras_dnase_demo_gr_hg19/hg38), a fragCounter read depth profile (read_depth_demo_hg38), and an RNA-seq counts matrix with condition labels (hmcl_counts_demo_dt_hg38, hmcl_conditions_demo_dt_hg38)

The raw files these were built from are in inst/extdata/.

Citation

To cite devgru in publications, use:

citation("devgru")

Contributing

Bug reports and feature requests are welcome on the issue tracker.