Skip to contents

Compute Motif-Based DMR Interactions

Source: R/dmrs_interaction.R
computeDMRsInteraction.Rd

Computes motif-based interactions between DMRs based on their motif similarity. Identifies pairs of DMRs with significant motif similarity and returns a data frame of interactions. Assigns directionality based on score, if available.

Usage

computeDMRsInteraction(
  dmrs,
  genome = "hg38",
  array = "450K",
  min_similarity = getOption("CMEnt.min_motif_similarity", 0.8),
  beta_locs = NULL,
  motif_site_flank_size = 5,
  find_components = TRUE,
  min_component_size = 2,
  query_components_with_jaspar = TRUE,
  plot_dir = NULL,
  output_prefix = NULL
)

Arguments

dmrs

Dataframe or GRanges object containing DMR coordinates and motif information

genome

Character. Genome version to use for sequence extraction (e.g., "hg38")

array

Character. Array platform type (e.g., "450K", "EPIC"). Must be NULL if input is not array-based (default: "450K")

min_similarity

Numeric. Minimum motifs PWM similarity threshold for considering DMRs are related (default: 0.8)

beta_locs

Data frame. Optional pre-computed genomic locations. If NULL, locations will be retrieved using getSortedGenomicLocs (default: NULL)

motif_site_flank_size

Integer. Number of base pairs to include as flanking regions around each site site (default: 5)

find_components

Logical. Whether to identify connected components of interacting DMRs (default: TRUE)

min_component_size

Integer. Minimum size of connected components to consider (default: 2)

query_components_with_jaspar

Logical. Whether to query connected components average PWMs against JASPAR database (default: TRUE)

plot_dir

Character. Directory to save diagnostic plots (optional). If NULL, no plots are saved (default: NULL)

output_prefix

Character. Prefix for output files to save interactions and components (optional). If NULL, results are not saved to file (default: NULL)

Value

A list with:

  • interactions: Data frame of motif-based DMR interactions with columns chr1, start1, end1, chr2, start2, end2, sim, and when components are requested, component_id

  • components: Data frame of discovered motif components

  • dmrs: The input DMRs in the same class as provided, with an added component_ids column containing comma-separated component IDs for each DMR (or NA when the DMR is not part of a retained component)

Examples

# Compute motif-based interactions for DMRs
dmrs <- data.frame(
    chr = c("chr16", "chr3"),
    start = c(53468112, 37459206),
    end = c(53468712, 37493431),
    start_site = c("cg00000029", "cg00000108"),
    start_seed = c("cg00000029", "cg00000108"),
    end_site = c("cg13426503", "cg08730726"),
    end_seed = c("cg13426503", "cg08730726"),
    seeds = c("cg00000029,cg13426503", "cg00000108,cg08730726")
)
dmrs_with_motifs <- extractDMRMotifs(dmrs, genome = "hg38", array = "450K")
interactions <- computeDMRsInteraction(
    dmrs_with_motifs,
    genome = "hg38",
    array = "450K",
    query_components_with_jaspar = FALSE
)