Manual
manual
User Manual:
Open the PDF directly: View PDF 
.
Page Count: 6
rGMAP
September 6, 2018
Type Package
Title Call hierarchical chromatin domains from HiC matrix by GMAP
Version 1.3.1
Date 08-23-2018
Author Wenbao Yu
Maintainer Wenbao Yu <yuw1@email.chop.edu>
Description Call hierarchical chromatin domains from HiC contact map by Gaussian Mix-
ture model And Proportion test
BugReports https://github.com/wbaopaul/rGMAP/issues
License GPL (>= 2)
LazyData TRUE
Imports data.table,
ggplot2,
mclust,
EMD,
caTools,
Matrix,
Rcpp (>= 0.12.5)
LinkingTo Rcpp
RoxygenNote 6.1.0
Suggests knitr,
rmarkdown
VignetteBuilder knitr
Rtopics documented:
data_simu.......................................... 2
hic_rao_IMR90_chr15 ................................... 2
plotdom ........................................... 3
rGMAP ........................................... 3
Index 6
1
2hic_rao_IMR90_chr15
data_simu generate simulated hic_mat and true tads
Description
generate simulated hic_mat and true tads
Usage
data_simu(stype = "poisson-dist", nratio = 2.5, mu0 = 200,
resl = 1)
Arguments
stype One of four types of simulated data in the manuscript: poission-dist, poission-
dist-hier, nb-dist, nb-dist-hier; poission- or nb- indicates poission distribution or
negative bionomial distribution -hier indicated subtads are generated nestly
nratio The effect size between intra- and inter domain, larger means higher intra-tad
contacts
mu0 The mean parameter, default 200
resl Resolution, default set to 1
Value
A list includes following elements:
hic_mat n by n contact matrix
hierTads True heirarchical domains
tads_true True TADs
hic_rao_IMR90_chr15
Normalized Hi-C data for IMR90, chr15 with resolution 10kb.
Description
Normalized Hi-C data for IMR90, chr15 with resolution 10kb.
Usage
hic_rao_IMR90_chr15
Format
A data table with 3 variables:
n1 bin 1
n2 bin 2
count normalized counts
plotdom 3
Source
Rao et al., Cell 2014, A 3D map of the human genome at kilobase resolution reveals principles of
chromatin looping
plotdom visualize hierarchical domains
Description
visualize hierarchical domains
Usage
plotdom(hic_dat, hiertads_gmap, start_bin, end_bin, cthr = 20,
resl = 10000)
Arguments
hic_dat hic contact matrix for a given chromosome, either a n by n matrix, or a 3 columns
data.frame <bin1> <bin2> <counts>
hiertads_gmap
the hierarchical domains called by GMAP
start_bin the start bin of the genome
end_bin the end bin of the genome
cthr the upper bound count threshold for color, default 20
resl reslution of Hi-C data, default 10000
rGMAP Detect hierarchical choromotin domains by GMAP
Description
Detect hierarchical choromotin domains by GMAP
Usage
rGMAP(hic_mat, index_file = NULL, resl = 10 *10^3, logt = T,
dom_order = 2, maxDistInBin = min(200, 2 *10^6/resl), min_d = 25,
max_d = 100, min_dp = 5, max_dp = 10, hthr = 0.95, t1thr = 0.5)
4rGMAP
Arguments
hic_mat • For single chromosome, supports three types of format:
–a 3-column Hi-C contact matrix, with columns the i_th, j_th bin of a
chromosom and the corresponding contact number
–a n by n matrix, with <i,j>th element corresponding to contact number
between the i_th and j_th bin of the chromosome
–a text file of the above two types of data
• For multiple chromosomes, a index_file indicates genomic coordinate for
each hic bin should be provided
index_file A 4-columns tab/space delimited text file indicates the genomic coordinates for
each bin (compatible with HiC-Pro); with columns bin_chr bin_start bin_end
bin_id
resl The resolution (bin size), default 10kb
logt Do log-transformation or not, default TRUE
dom_order Maximum level of hierarchical structures, default 2 (call TADs and subTADs)
maxDistInBin Only consider contact whose distance is not greater than maxDistInBIn bins,
default 200 bins (or 2Mb)
min_d The minimum d (d: window size), default 25
max_d The maximum d (d: window size), default 100
min_dp The minmum dp (dp: lower bound of tad size), defalt 5
max_dp The maximum dp (dp: lower bound of tad size), defalt 10. min_d, max_d,
min_dp and max_dp should be specified in number of bins
hthr The lower bound cutoff for posterior probability, default 0.95
t1thr Lower bound for t1 for calling TAD, default 0.5 quantile of test statistics of
TADs, 0.9 of subTADs
Value
A list includes following elements:
tads A data frame with columns start, end indicates the start and end coordinates of
each domain, respectively
hierTads A data frame with columns start, end, dom_order, where dom_order indicates
the hierarchical status of a domain, 1 indicates tads, 2 indicates subtads, and so
on
params A data frame gives the final parameters for calling TADs
Examples
## On simulated data:
library(rGMAP)
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)$hierTads
true_domains
predicted_domains
## On an real data example
rGMAP 5
hic_rao_IMR90_chr15 # normalized Hi-C data for IMR90, chr15 with resolution 10kb
res = rGMAP(hic_rao_IMR90_chr15, resl = 10 *1000, dom_order = 2)
names(res)
#quickly visualize some hierarchical domains
pp = plotdom(hic_rao_IMR90_chr15, res$hierTads, 6000, 7000, 30, 10)
pp$p2
