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

Description Usage Arguments Value Examples

View source: R/rGMAP_source.R

Detect hierarchical choromotin domains by GMAP

1
2
3

rGMAP(hic_obj, index_obj = 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)

`hic_obj`	HiC contact matrix, generally either a 3-column data frame/data table or a tab/space delimited text file, corresponding to the i_th, j_th bin of a chromosome and the contact frequency For single chromosome, could also be a n by n matrix For multiple chromosomes, an index_obj indicates genomic coordinate for each hic bin should be provided
`index_obj`	A 4-columns tab/space delimited text file (or corresponding data frame/data table) indicates the genomic coordinates for each bin (compatible with HiC-Pro); with columns bin_chr bin_start bin_end bin_id, default NULL; when index _obj was given, it should be consitent with hic_obj
`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

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

## 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
hic_rao_IMR90_chr15   # normalized Hi-C data for IMR90, chr15 with resolution 10kb
res = rGMAP(hic_rao_IMR90_chr15, NULL, resl = 10 * 1000, dom_order = 2)
names(res)
#quickly visualize some hierarchical domains
pp = plotdom(hic_rao_IMR90_chr15, NULL, res$hierTads, NULL, 6000, 7000, 30, 10)
pp$p2