Skip to content

Latest commit

 

History

History
191 lines (141 loc) · 7.2 KB

TF_signal_around_G4s.md

File metadata and controls

191 lines (141 loc) · 7.2 KB
Raw

TF ChIP-seq signal at endogenous G4s

Plot the TF ChIP-seq signal around endogenous G4s (BG4 ChIP-seq peaks) and control sites (potential G4s in open chromatin promoters that do not form endogenous G4s). In addition, investigate the distribution of TF binding sites around endogenous G4s.

Generate BigWig files

Obtain one TF ChIP-seq bam-file from ENCODE for each of the selected experiments in K562

library(tidyverse)

# ====== Load meta data
ENCODE_K562_meta <- read_tsv(file = "ENCODE_K562_Jan2019_all_Meta.tsv", col_names = T)
BED_considered <- read_csv(file ="GAT_analysis/GAT_K562_BG4_100Peaks_Jan2019_Gat-Analysis_trim.csv")

# keep bam files from hg19 that were used for considered Bed files
# this still contains multiple replicates and potentially archived files; prefer low number replicates and "released" over "archived"; keep only one bam file each

ENC_filt <-  ENCODE_K562_meta %>% filter(`File format`=='bam', `Output type`=='alignments', Assembly=='hg19', `Experiment accession` %in%  BED_considered$Experiment.accession) %>%
  arrange(`Experiment accession`, desc(`File Status`), `Biological replicate(s)`) %>% 
  group_by(`Experiment accession`) %>% slice(1)  %>% ungroup() %>% 
  select(`File download URL`) %>%
  write_tsv("bam_download/20201220_LINKS_K562_bamfiles_1replicate.txt", col_names = F)

Download files

sbatch --mem 4G -o DL.out -e DL.err -J bam_DL --wrap "xargs -n 1 curl -O -L < 20201220_LINKS_K562_bamfiles_1replicate.txt"

Generate bigWig-files normalized to sequencing depth generated using the bamCoverage function in deeptools

# Index bam
for FILE in *.bam
do
sbatch -o index.%j.log -J BamIndex --mem=5G \
--wrap "samtools index $FILE"
done

# Coverage files at 10 nt bin size
for FILE in *.bam
do
bname=`basename $FILE .bam`
sbatch -o log/%j.$f.log -J BCov --mem=8G \
--wrap "bamCoverage -b $FILE \
-o BigWig_10nt/$bname.bs10.bl.RPKM.bw \
--binSize 10 \
--blackListFileName reference_data/hg19/hg19.blacklist_merge1000nt.bed \
--normalizeUsing RPKM"
done

Signal plots

Generate signal profile plots 400bp around BG4 ChIP-seq peak centers and respective control regions (potential G4s in open chromatin promoters that do not form endogenous G4s) using functions computeMatrix and plotProfile.

cd BigWig_10nt

for f in *.bs10.bl.RPKM.bw
do
sbatch -o logs/${f%%.bs10.bl.RPKM.bw}.log -J DT_Profile --mem=5G \
--wrap "computeMatrix reference-point \
--referencePoint center \
-b 400 -a 400 \
-S \
$f \
-R \
reference_data/Quadruplex/K562_async_rep1-3.mult.5of8-OQS_Stranded.bed \
reference_data/openOQs_noBG4_Stranded_1kbupTSS.bed \
--skipZeros \
-o ../ChIP_Profile/ENCODE2019_bam_1rep_2020/${f%%.bs10.bl.RPKM.bw}.mat.gz && 
plotProfile -m ../ChIP_Profile/ENCODE2019_bam_1rep_2020/${f%%.bs10.bl.RPKM.bw}.mat.gz \
-out ../ChIP_Profile/ENCODE2019_bam_1rep_2020/${f%%.bs10.bl.RPKM.bw}.pdf \
--refPointLabel PeakCenter \
--regionsLabel G4 OQ \
--outFileNameData ../ChIP_Profile/ENCODE2019_bam_1rep_2020/${f%%.bs10.bl.RPKM.bw}_profile.tab \
--colors green grey \
--yMin 0 \
--plotHeight 6 \
--plotWidth 6" 
done

Distribution of TF signal maxima around BG4 ChIP-seq peaks

Binned average signal intensity for each TF experiment are saved int *.tab. Extract the information on signal around BG4 ChIP-seq sites (3rd line in each tab file):

cd ../ChIP_Profile/ENCODE2019_bam_1rep_2020/

for f in *.tab
do
bname=`basename $f _profile.tab`
sed -n 3p $f > ../../Signal_around_BG4/$bname.tab
done

Plot data in using a custom R script

library(tidyverse)
library(readr)
library(ggsignif)
library(ggrepel)

#################################
# <<<<<<  G4s  >>>>>>>>>
#############################
# Fetch data
#generate master file of average TF signal around BG4 peaks
Avr_Signal_at_G4 <- list.files(path="Signal_around_BG4/", full.names = TRUE) %>% lapply(read_tsv) %>% bind_rows 

# reformat and keep relevant columns
Avr_Signal_at_G4 <- Avr_Signal_at_G4 %>% separate(col = "bins", into = c("ENCODE_bam_ID", "B"), sep = ".b")  %>% select(-2, -3)

# find  position and value of maximum count (there are 1-80 '10nt'-bins representing the -400bp to +400bp window)
Max_signal <- Avr_Signal_at_G4 %>% 
  pivot_longer(`1`:`80`, names_to = "position", values_to = "cnt") %>% 
  group_by(ENCODE_bam_ID) %>% 
  slice(which.max(cnt)) %>%
  mutate(position = as.integer(position)*10 -400)    # col1-80 represent -400 to 400 bp; therefore, pos = x*10 -400


# Integrate information on genomic enrichment from GAT analysis 
ENCODE_K562_meta_bam <- read_tsv(file = "ENCODE_K562_Jan2019_BAM_Meta.tsv", col_names = T)
GAT_K562 <- read_csv(file = "GAT_K562_BG4_100Peaks_Jan2019_Gat-Analysis_trim.csv", col_names = T)

Max_signal$ENCODE_exp_ID <- ENCODE_K562_meta_bam$`Experiment accession`[match(Max_signal$ENCODE_bam_ID, ENCODE_K562_meta_bam$`File accession`)]
Max_signal$target <- GAT_K562$Experiment.target[match(Max_signal$ENCODE_exp_ID, GAT_K562$Experiment.accession)]
Max_signal$GAT_DHS <- GAT_K562$DHS_fold[match(Max_signal$ENCODE_exp_ID, GAT_K562$Experiment.accession)]


##################
#### Plot data ###
##################

Signal_Plot <- Max_signal

# Visual inspection suggests that there are a few noisy, low intensity data sets 
# these can easily filtered using a cutoff for intensity
Signal_Plot <- Signal_Plot %>% filter(cnt > 5)


# Remove histone marks and PolR2
Signal_Plot <- Signal_Plot %>% filter(!grepl('H2A|H3|H4|POLR', target))

# Remove  protein-tag information  "eGFP", "3xFLAG" etc
Signal_Plot  <- Signal_Plot %>% mutate(target = str_remove_all(target, "eGFP-|3xFLAG-"))

# Highlight TFs within a 20 bp window around center
Signal_Plot <-  Signal_Plot %>% mutate( In_20bp_window = ifelse(position >= -20 & position <=20, "blue", "black"))

# check how many TFs bind within a 20, 40 or 60 bp window
Signal_Plot %>% filter(position == 0 ) %>% nrow() #7
Signal_Plot %>% filter(position >= -10 & position <=10) %>% nrow() #40
Signal_Plot %>% filter(position >= -20 & position <=20) %>% nrow() #100
Signal_Plot %>% filter(position >= -30 & position <=30) %>% nrow() #148
Signal_Plot %>% filter(position >= -40 & position <=40) %>% nrow() #177

# remove TFs that don't have a maximum inside the explored 400bp interval
Signal_Plot  <- Signal_Plot %>% filter(position > -400 & position < 400) 

# Plot G4 association
gg_G4s_OQS <- ggplot(Signal_Plot, aes(x=position, y=GAT_DHS)) + 
  geom_point(color=Signal_Plot$In_20bp_window) + 
  scale_color_gradient2(midpoint = 20, low = "blue", mid = "black", high = "red", space = "Lab" ) +
  theme_minimal() +
  labs(title="TF position vs GAT DHS", y="enrichment at endogenous G4s", x="TF signal maximum relative to G4") +
  geom_label_repel(aes(label=ifelse(
    (GAT_DHS > 5 & position >= -20 & position <= 20), as.character(target),'')),
    box.padding = 0.4, point.padding = 0.4, size=rel(2.5), segment.color = 'grey80', color=Signal_Plot$In_20bp_window) +
  theme(panel.background = element_rect(fill = "transparent"), 
        plot.background = element_rect(fill = "transparent", color = NA))

plot(gg_G4s_OQS) 

ggsave("figures/TFPosition_G4_400bp.pdf", width = 17/2.54, height = 14/2.54, useDingbats=F)


``