GGR UCSC trackhub

This trackhub includes the following supertracks:

• GGR_ChIP_seq_diff_peaks
• GGR_ChIP_seq_peak_mean_differential
• GGR_DNase_seq_diff_peaks
• GGR_DNase_seq_peak_mean_differential
• GGR_RNA_seq_degs
• GGR_RNA_seq_mean_signal
• GGR_Hi_C_tads
• GGR_Hi_C_directionality_index

ChIP-seq

Factors analyzed:

• GR, BCL3, CEBPB, EP300, FOSL2, HES2, JunB and cJun.

Time points:

• 0, 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 10 and 12 hours of the GC response.

• ChIP-seq samples were processed with Duke-GCB CWL ChIP-seq pipeline. Here is a list of the most relevant steps:

  • Sequencing by Illumina HiSeq
  • Raw 50-bp paired-end reads were trimmed using Trimmomatic (v.0.32) (Bolger et al. 2014; doi:10.1093/bioinformatics/btu170).
  • Trimmed reads were mapped to human reference genome assembly GRCh38 using bowtie (v.0.12.9) (Li and Durbin 2009; doi:10.1093/bioinformatics/btp324) with options -m 1 -v 2 --best --strata
  • PCR duplicates were filtered using Picard tools, http://picard.sourceforge.net
  • ENCODE blacklist regions were filtered using bedtools (Quilan and Hall 2010; doi: 10.1093/bioinformatics/btq033)
  • Mappings were normalized by RPKM as implemented in deepTools after extending reads 200 bp in the 3' direction and summing counts in 50-bp bins (Ramirez et al. 2014; doi: 10.1093/nar/gku365).
  • Read mappings were scaled by a constant factor divided by the total number of mapped reads
  • Mean of read mappings was taken across replicates and within time points.

GGR_ChIP_seq_diff_peaks

This track contains all differential (and non-differential) binding peaks colored by their direction (increased, decreased, or increased and decreased binding) and strength of differential binding (FDR = 0.01, 0.05 0.1 0.2) across the time course. Differential binding analyses were performed using the negative binomial model implemented in edgeR (Robinson et al. 2010; doi: 10.1093/bioinformatics/btp616). We controlled for batch effects by incorporating all significant surrogate variables estimated using SVAseq (Leek 2014; doi: 10.1093/nar/gku864) into the model as covariates. Normalization factors were computed based on total mapped library size.

GGR_ChIP_seq_peak_mean_differential

For each factor, this supertrack is composed of:

• ChIP_mean_$FACTOR_0_hr. Signal reference track. Contains:
  • ChIP_mean_$FACTOR_0_hr_mean. Mean signal file (bigWig) at time point 0, created using the procedure described above.
  • ChIP_mean_$FACTOR_0_hr_union_peaks. Peaks indicated by stacked bedGraph. Peaks were called within-replicate using macs2 (v.2.1.0.20151222)(Zhang et al. 2008; doi: 10.1186/gb-2008-9-9-r137) with parameters --nomodel --extsize $EXTSIZE -g hs -q 0.05 where $EXTSIZE was the fragment length estimated by SPP (v.2.0) (Karchenko et al. 2008; doi:10.1038/nbt.1508). The peaks were then merged across replicates to yield a union set, which is the set shown in this hub.
• ChIP_mean_${FACTOR}_vs_t00. Difference in signal along time with respect to time point 0.
  • $FACTOR.$TIMEPOINTvst00: For each $FACTOR and $TIMEPOINT the signal was subtracted using the deepTools (Ramirez et al. 2014; doi: 10.1093/nar/gku365) subcommand bamCompare and normalized using the SES method (Diaz et al. 2012; doi: 10.1515/1544-6115.1750), parameters --ratio subtract --sampleLength=10000 --numberOfSamples=1000000. The resulting bigWigs have positive and negative values which represent respectively an increased and decreased signal compared with time point 0, non-DEX.

DNase-seq

• DNase-seq samples were processed with Duke-GCB CWL DNase-seq pipeline. Here is a list of the most relevant steps:
  • Sequencing by Illumina HiSeq
  • After stripping custom barcodes, raw reads were mapped with bowtie (v.0.12.9) (Langmead et al. 2009; doi:10.1186/gb-2009-10-3-r25) with options --trim3 30 --seedlen 20 --seedmms 1 -m 1 --best --strata.
  • Mappings were filtered for ENCODE blacklisted regions.
  • Mappings were filtered for PCR artifacts using a custom python script windowTrimmer.py.
  • ENCODE blacklist regions were filtered using bedtools (Quilan and Hall 2010; doi: 10.1093/bioinformatics/btq033)
  • Mappings were normalized by RPKM as implemented in deepTools after extending reads 200 bp in the 3' direction and summing counts in 50-bp bins (Ramirez et al. 2014; doi: 10.1093/nar/gku365).
  • Read mappings were scaled by a constant factor divided by the total number of mapped reads
  • Mean of read mappings was taken across replicates and within time points.

GGR_DNase_seq_diff_peaks

This track contains all differential (and non-differential) DHSs colored by their direction (increased, decreased, or increased and decreased accessibility) and strength of differential accessibility (FDR = 0.01, 0.05 0.1 0.2) across the time course.

GGR_DNase_seq_peak_mean_differential

This supertrack is composed of:

• DNase_mean_0_hr. Signal reference track. Contains:
  • DHSs_peak_t00_mean. Mean signal file (bigWig) at time point 0, created using the procedure described above.
  • DHSs_$TRAJECTORY_FDR_$FDR. DHSs colored by their direction (increased, decreased, or increased and decreased accessibility) and strength of differential accessibility (FDR = 0.01, 0.05 0.1 0.2) across the time course. The accessibility is defined as follows:
    • increased accesibility or opening DHSs: logFC over pre-DEX levels > 0 in 90% of the time points.
    • decreased accesibility or closing DHSs: logFC over pre-DEX levels < 0 in 90% of the time points.
    • increased and decreased accesibility or ambiguous DHSs: others.
• DNaseI_mean_vs_t00. Difference in signal along time with respect to time point 0.
  • dnase.{TIMEPOINT}vst00: For each $TIMEPOINT the signal was subtracted using the deepTools (Ramirez et al. 2014; doi: 10.1093/nar/gku365) subcommand bamCompare --ratio subtract and normalized using the SES method (Diaz et al. 2012; doi: 10.1515/1544-6115.1750) with parameters --sampleLength=10000 and --numberOfSamples=1000000. The resulting bigWigs have positive and negative values which represent respectively an increased and decreased signal compared with time point 0, non-DEX.

RNA-seq

• RNA-seq samples were processed with Duke-GCB CWL RNA-seq pipeline. Here is a list of the most relevant steps:

• Sequencing by Illumina HiSeq

• Raw 50-bp paired-end reads were trimmed using Trimmomatic (v.0.32) (Bolger et al. 2014; doi:10.1093/bioinformatics/btu170).
• Trimmed reads were mapped to human reference genome assembly GRCh38 using STAR aligner (v.2.4.1a) (Dobin et al. 2012: doi:10.1093/bioinformatics/bts635) using the 2-pass procedure wherein novel splice junctions discovered on a first pass of alignments are compiled into a splice junction database used to inform a second pass of alignments as previously described (Engstrom et al. 2013; doi:10.1038/nmeth.2722). STAR aligner was run using default setting except with filtering for non-canonical novel intron junctions and with "sjdbOverhang" parameter set to read length - 1 (49).
• Mappings were normalized by RPKM as implemented in deepTools (Ramirez et al. 2014; doi: 10.1093/nar/gku365). (Ramirez et al. 2014; doi: 10.1093/nar/gku365).
• RNA-seq read counts were quantified in gene body exons as annotated in GENCODE (v.22) (Harrow et al. 2012; doi: 10.1101/gr.135350.111) using featureCounts (v.1.4.6-p4) (Liao et al. 2014; doi: 10.1093/bioinformatics/btt656) where both paired reads were required to map to the same gene to be counted. Differential expression analysis was performed in the same manner as the differential accessibility analysis.

[GGR_RNA_seq_mean_signal

For each time point and strand, the mean signal tracks contains:

• GGR_RNA_seq_mean_signal. Signal tracks for positive (plus) and negative (minus) strands.
• GGR_RNA_seq_DEGs. Differentially expressed genes (DEGs) colored by their direction (increased, decreased, or increased and decreased expression) and strength of differential expression (FDR = 0.01, 0.05 0.1 0.2) across the time course.

GGR_RNA_seq_degs

This track contains all DEGs (and tested non-DEGs) colored by their direction (increased, decreased, or increased and decreased expression) and strength of differential expression (FDR = 0.01, 0.05 0.1 0.2) across the time course.

Hi-C

• Sequenced by Illumina HiSeq4000

GGR_Hi_C_tads

• Trackhub contains TADs at 5kb, 10kb and 25kb resolutions computed with arrowhead from Juicer v1.5 (Durand et al. 2016; doi: 10.1016/j.cels.2016.07.002).

GGR_Hi_C_directionality_index

• Trackhub contains Directionality Index as calculated in Dixon et al. 2012 (doi:10.1038/nature11082)