overlap2hm()

Adrià Mitjavila Ventura

13 May, 2021

Default overlap from ChromHMM

knitr::include_graphics("../demodata/conditionX_8_overlap.png")

Run chromHMMviewR::overlap2hm()

Install and load

# install devtools if not installed
if(!require(devtools)){ install.packages("devtools") }

# install chromHMMviewR if not installed
if(!require(chromHMMviewR)){ devtools::install_github("amitjavilaventura/chromHMMviewR") }

# load chromHMMviewR
library(chromHMMviewR)

Read and explore demodata

# read overlap file from the demodata
overlaps <- read.delim("../demodata/conditionX_8_overlap.txt")

# head
head(overlaps)

##   state..Emission.order. Genome.. CpGIsland.hg38.bed.gz RefSeqExon.hg38.bed.gz
## 1                      1  0.22264              70.46286                7.98271
## 2                      2  0.44664              13.64083                3.77196
## 3                      3  0.42034              68.67053                9.98436
## 4                      4  9.92537               1.04092                2.02856
## 5                      5  0.88475               0.53824                1.40042
## 6                      6  2.47907               1.48283                1.31637
##   RefSeqGene.hg38.bed.gz RefSeqTES.hg38.bed.gz RefSeqTSS.hg38.bed.gz
## 1                1.42469               1.48807              66.75551
## 2                1.57501               4.35027               9.67735
## 3                1.85742               2.98860              66.57028
## 4                1.49961               2.29835               0.82918
## 5                1.44084               2.46524               1.49866
## 6                1.41083               1.57865               1.75210
##   RefSeqTSS2kb.hg38.bed.gz
## 1                 22.97113
## 2                 22.44641
## 3                 27.08225
## 4                  1.41808
## 5                  2.96553
## 6                  3.78615

Call overlap2hm()

Minimum

# set the states with numbers
# states should be named later depending on the results
states <- paste("S", 1:8, sep = "")

# set the names of the regions in the same order as in the overlap file
# (they do not need to have the exact same name as in the file)
# in this case, the regions are the default from ChromHMM hg38
regions <- c("% Genome", "CpG islands", "Exon RefSeq", "Gene RefSeq", "TES RefSeq", "TSS RefSeq", "TSS +/- 2kb RefSeq")

# call overlap2hm()
overlap2hm(data = overlaps, states = states, regions = regions)

## Loading required package: dplyr

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

## Loading required package: magrittr

## Loading required package: reshape2

## Loading required package: ggplot2

## Loading required package: ggpubr

## Using State as id variables

From a file

# call overlap2hm()
overlap2hm(data = "../demodata/conditionX_8_overlap.txt", states = states, regions = regions)

## Using State as id variables

Customize the output

# change title and subtitle
overlap2hm(data = overlaps, states = states, regions = regions, title = "This is a title", subtitle = "This is a subtitle")

## Using State as id variables

# change color of the heatmap: color = "gray30"
overlap2hm(data = overlaps, states = states, regions = regions, color = "gray30")

## Using State as id variables

#if(!require(patchwork)){ devtools::install_github("thomasp85/patchwork") }
library(patchwork) # to show the plots side-by-side

# change the scaling of the color (by default, chromHMM normalizes each column separately)
# must be one of 'scale' (default), 'percent' or 'value'
# the numbers showing the enrichment don't change
overlap2hm(data = overlaps, states = states, regions = regions, 
           scale_color = "scale", title = "'scale'") + # scale by column (default in ChromHMM and chromHMMviewR)
  
overlap2hm(data = overlaps, states = states, regions = regions, 
           scale_color = "percent", title = "'percent'") + # percentage of each cell within its column 

overlap2hm(data = overlaps, states = states, regions = regions, 
           scale_color = "value", title = "'value'") # use the enrichment value in each cell

## Using State as id variables
## Using State as id variables
## Using State as id variables

# change the size of the likelihood labels (default: score_size = 2)
overlap2hm(data = overlaps, states = states, regions = regions, score_size = 3)

## Using State as id variables

# without likelihood labels: show_score = F
overlap2hm(data = overlaps, states = states, regions = regions, show_score = F)

## Using State as id variables