TADpole: Call hierarchical TADs

In 3DGenomes/TADpole: Call hierarchical TADs

Description

Computes a constrained hierarchical clustering of genomic regions in a HiC experiment, choosing the optimal amount of information from the HiC matrix and selecting the most informative number of TADs.

Usage

TADpole(
  mat_file,
  max_pcs = 200,
  min_clusters = 2,
  bad_frac = 0.01,
  chr,
  start,
  end,
  resol,
  centromere_search = FALSE
)

Arguments

mat_file	path to the input file. Must be in a tab-delimited matrix format.
max_pcs	The maximum number of principal components to retain for the analysis.
min_clusters	Minimum number of clusters into which partition the chromosome.
bad_frac	fraction of the matrix to falg as bad rows/columns.
chr	string with the chromosome name.
start	numeric start position of the region or of the chromosome.
end	numeric end position of the region or of the chromosome.
resol	numeric resolution/binning of the Hi-C experiment.
centromere_search	split the matrix by the centrormere into two smaller matrices representing the chromosomal arms. Useful when woring with big (>15000 bins) datasets.

Details

The 'centromere_search' parameter will split the matrix into two by the region with the longes stretch of bad (low signal) rows/columns. It will do so regardless of whether this stretch represents a true centromere or not. Note that this feature is useful when processing an entire chromosome, but be cautious of interpreting the partitions as the two chromosomal arms (p and q) when working with smaller regions.

Value

'tadpole' object that defines the clustering of genomic regions.

Examples

mat_file <- system.file("inst/extdata", "raw_chr18:460-606_20kb.tsv", package = "TADpole")
tadpole <- TADpole(mat_file, chr = "chr18", start = 496, end = 606, resol = 20000)

3DGenomes/TADpole documentation built on Jan. 30, 2020, 8:17 p.m.