SCC

Sum Of Chromatin State by Contact

Step 1

Get contact scores for every region with a chromHMM label.

# python 1_get_scores_mcool.py <mcool file with res> <ChromHMM bedfile> <outfile> [margin size]

python 1_get_scores_mcool.py 4DNFI9E222YJ.mcool::/resolutions/1000 \
							 endoderm_ChromHMM.bed.gz \
							 all_loops.bedpe.gz

Step 2

Get labels of every contact region found in step 1. In the multi-cell type analysis we also merge cell types before this step, see preprocess_scripts.

# chromHMM_total.bed.gz
#  chr1    10000    177200    Quies    0    .    10000    177200    255,255,255
#  chr1    257849    297849    Quies    0    .    257849    297849    255,255,255
#  chr1    586020    777820    Quies    0    .    586020    777820    255,255,255

# all_loops.bedpe.gz
#  chr1    778000    779000    chr1_endo    777820    778420    0.17871180991657024
#  chr1    779000    780000    chr1_endo    777820    778420    0.23094778295675608
#  chr1    787000    788000    chr1_endo    777820    778420    0.05381787728341296

./2_get_contact_labels.sh all_loops.bedpe.gz chromHMM_total.bed.gz

Step 3

Create a summed matrix of contacts from previous step

# include path to scored bedfiles within script (created in step 2)
# 	state_bedfiles/scored/*bed

./3_create_matrix.sh

The resulting matrix_connections.tsv is ready for clustering.

Supplemental Data for the manuscript

Supplemental_data_S1: SCC matrix resulting from the sum of chromatin state by contact

Supplemental_data_S2: cluster membership for each row resulting from k-means clustering on SCC matrix

Supplemental_table_S1: differential expression analysis