In wbaopaul/rGMAP: Call hierarchical chromatin domains from HiC matrix by GMAP

knitr::opts_chunk$set(
  collapse = TRUE,
  warning = F, message = F,
  comment = "#>"
)

Introduction

• rGMAP is an R package for implementing GMAP algorithm for detecting topologically associating domains (TADs) and sub-domains (sub-TADs)
• The detailed of the algorithm can be found in Yu, W., He, B., & Tan, K. (2017). Identifying topologically associating domains and subdomains by Gaussian Mixture model And Proportion test. Nature Communications, 8, 535

Updates

• The inputs of rGMAP function support both data frame/data table and tab/space delimited text file for both hic_obj and index_obj, and are consitent with the inputs of plotdom function
• n by n matrix of hic_obj is still supported, where n is the total number of bins of a hic contact map for a single chromosome

Note

• The index file index_obj was required in multiple chromosome mode
• If the input hic_obj is just from a sigle chromosome and no index_file was provided: the first two columns should be the i_th, j_th bin of the chromosome if hic_obj is in 3-column format (i j count)

Installation

Install from Github

library(devtools)
install_github('wbaopaul/rGMAP')

Install from source codes

install.packages('path to rGMAP_1.4.tar.gz', repos = NULL, type = 'source')

Examples

Work on an published Hi-C data

Data from Rao et al., Cell 2014, A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping

KR normalized Hi-C data for IMR90 cell line, on chr15 with resolution 10kb

library(rGMAP)
hic_rao_IMR90_chr15   
res = rGMAP(hic_rao_IMR90_chr15, index_obj = NULL, resl = 10 * 1000, dom_order = 2)
names(res)

#quickly visualize some hierarchical domains
pp = plotdom(hic_rao_IMR90_chr15, index_obj = NULL, res$hierTads, chr0 = NULL, 5950, 6900, 30, 10000)
pp$p2

Compatible with HiC-Pro output matrix and index file

• The format of the index file (10kb for example):

chr1  0 10000   1
chr1    10000   20000   2
chr1    20000   30000   3
chr1    30000   40000   4
......

• The format of the hic_mat data:

10  127 10.372577
10  188 4.161271
10  281 36.072713
17  5822    2.418056
......

Work on an simulated example

set.seed(1)
simu_res = data_simu('poisson-dist-hier')
true_domains = simu_res$hierTads
simu_mat = simu_res$hic_mat
predicted_domains = rGMAP(simu_mat, resl = 1)$hierTads

pp = plotdom(simu_mat, NULL, predicted_domains, NULL, 1, 1000, 20, resl = 1)
pp$p2

#true_domains
#predicted_domains

Session Information

devtools::session_info()

wbaopaul/rGMAP documentation built on Nov. 18, 2020, 9:37 p.m.