Calling significant interactions from Capture-C or Capture-HiC

Usage

interaction_calling_ChiCMaxima.py -bed break.20copy_hg19.bed -bdg 20copy_hg19.captureC.bdg.bed -o 20copy_hg19.called_interactions

Parameters

Parameter tuning guide is here: https://static-content.springer.com/esm/art%3A10.1186%2Fs13059-019-1706-3/MediaObjects/13059_2019_1706_MOESM3_ESM.pdf

4 parameters to consider: w=20, s=0.05, b=20 kb (fixed in the new ChiCmaxima version), c=1.5 Mb.

Usually parameter tunning is recommended, but there is no rules for which directions should be.

Notes

Another tool for captureC analysis: CaptureCompendium: a comprehensive toolkit for 3C analysis. This is actually the same lab who developed captureC.

ChiCMaxima_Caller has to roll back to v0.9, because the latest one does not have the enrichment score.

w=$1
s=$2
c=$3

# -w/--window_size [LOCAL MAXIMUM CALLING WINDOW] -s/--loess_span [LOESS SPAN] -c/--cis_window [GENOMIC SEPARATION THRESHOLD]
# c('output','o', 1, 'character',"output.ibed"),
# c('window_size', 'w', 1, 'integer', 20),
# c('loess_span','s',1,'numeric',0.05),
# c('input','i',1,'character',"chr1_mESCs.ibed"),
# c('cis_window','c',1,'integer',1500000),

bdg_to_ibed.py merged.non_target.bdg target.merge.bed
ChiCMaxima_Caller.r -i input.ibed -o 20copy -w $w -s $s -c $c
sed -i '1d' 20copy_interactions.ibed
awk -F"\t" '$12 >= 1 { print $2"\t"$3"\t"$4"\t"$7"\t"$8"\t"$9"\t"$12 }' 20copy_interactions.ibed > 20copy_interactions.mango
cut -f 4,5,6,7 20copy_interactions.mango > 20copy_interactions.oe.bed
bedtools intersect -a 20copy_interactions.oe.bed -b CTCF.3kb.bed -u | cut -f 4 > A.list
bedtools intersect -a 20copy_interactions.oe.bed -b CTCF.3kb.bed -v | cut -f 4 > B.list
boxplot_two_list.py -l1 A.list -l2 B.list

# create_tracks.py --current_dir

Review on ChiCMaxima

  1. https://www.frontiersin.org/articles/10.3389/fgene.2022.786501/full

CHiCAGO and CHiCMaxima can be a good compromise between false positive and false negative rates, as these identify 4–18 times as many interactions as CHiCANE, and the proportion of those significant interactions that overlap with regulatory features is not much below CHiCANE’s.

  1. A supervised learning framework for chromatin loop detection in genome-wide contact maps

ChiCMaxima avoids statistical tests by using strategies from the signal processing field to find local maxima and integrates biological replicate information to reduce false-positive rates

  1. Detecting chromosomal interactions in Capture Hi-C data with CHiCAGO and companion tools

CHiCMaxima43 proposes a ranking approach for interaction detection that does not involve a formal statistical test. It averages through the interaction profile with a sliding window and reports local maxima, which occur at a frequency above local averaged count, as true interactions.

Better to use CTCF or H3K27ac peak to determine the best cutoff.

captureC data analysis

https://www.nature.com/articles/s41467-020-20809-6#Sec11

code @ github.