all barplots generated for ps and lig
This commit is contained in:
parent
0220960975
commit
3f8d6695a4
5 changed files with 656 additions and 0 deletions
265
scripts/plotting/barplots_subcolours_aa_LIG.R
Executable file
265
scripts/plotting/barplots_subcolours_aa_LIG.R
Executable file
|
|
@ -0,0 +1,265 @@
|
|||
#!/usr/bin/env Rscript
|
||||
getwd()
|
||||
setwd("~/git/LSHTM_analysis/scripts/plotting")
|
||||
getwd()
|
||||
|
||||
#########################################################
|
||||
# TASK: output barplot by position with each bar coloured by
|
||||
# its stability value and active site positions indicated
|
||||
# according to colour specified in "subcols_axis_PS.R"
|
||||
#########################################################
|
||||
|
||||
#=======================================================================
|
||||
|
||||
############################################################
|
||||
# 1: Installing and loading required packages and functions
|
||||
############################################################
|
||||
|
||||
#source("Header_TT.R")
|
||||
library(ggplot2)
|
||||
library(data.table)
|
||||
source("barplot_colour_function.R")
|
||||
#source("subcols_axis.R")
|
||||
source("subcols_axis_PS.R")
|
||||
|
||||
# should return the following dfs, directories and variables
|
||||
# mut_pos_cols
|
||||
# mut_pos_cols_lig
|
||||
# my_df_cols
|
||||
# my_df_u_cols
|
||||
# my_df_u_lig_cols
|
||||
# dup_muts_cols
|
||||
|
||||
cat(paste0("Directories imported:"
|
||||
, "\ndatadir:", datadir
|
||||
, "\nindir:", indir
|
||||
, "\noutdir:", outdir
|
||||
, "\nplotdir:", plotdir))
|
||||
|
||||
cat(paste0("Variables imported:"
|
||||
, "\ndrug:", drug
|
||||
, "\ngene:", gene
|
||||
, "\ngene_match:", gene_match
|
||||
, "\nLength of upos:", length(upos)
|
||||
, "\nAngstrom symbol:", angstroms_symbol))
|
||||
|
||||
# clear excess variable
|
||||
rm(dup_muts_cols, mut_pos_cols, my_df_cols, my_df_u_cols, upos)
|
||||
|
||||
#=======================================================================
|
||||
# !!! very important!!!!
|
||||
#================
|
||||
# Inspecting mut_pos_cols
|
||||
# position numbers and colours and assigning axis colours based on lab_fg
|
||||
# of the correct df
|
||||
# open file from desktop ("sample_axis_cols") for cross checking
|
||||
#================
|
||||
if ( nrow(mut_pos_cols_lig) == length(unique(my_df_u_cols_lig$position)) ){
|
||||
print("PASS: lengths checked, assigning axis colours")
|
||||
my_axis_colours = mut_pos_cols_lig$lab_fg
|
||||
cat("length of axis colours:", length(my_axis_colours)
|
||||
, "\nwhich corresponds to the number of positions on the x-axis of the plot")
|
||||
}else{
|
||||
print("FAIL:lengths mismatch, could not assign axis colours")
|
||||
quit()
|
||||
}
|
||||
|
||||
# further sanity checks
|
||||
table(mut_pos_cols_lig$lab_bg)
|
||||
check_lab_bg = sum( table(mut_pos_cols_lig$lab_bg) ) == nrow(mut_pos_cols_lig) # should be True
|
||||
check_lab_bg
|
||||
|
||||
table(mut_pos_cols_lig$lab_bg2)
|
||||
check_lab_bg2 = sum( table(mut_pos_cols_lig$lab_bg2) ) == nrow(mut_pos_cols_lig) # should be True
|
||||
check_lab_bg2
|
||||
|
||||
table(mut_pos_cols_lig$lab_fg)
|
||||
check_lab_fg = sum( table(mut_pos_cols_lig$lab_fg) ) == nrow(mut_pos_cols_lig) # should be True
|
||||
check_lab_fg
|
||||
|
||||
# sanity checks:
|
||||
if (check_lab_bg && check_lab_bg2 && check_lab_fg) {
|
||||
print("PASS: No. of assigned colours match length")
|
||||
}else{
|
||||
print("FAIL: length of assigned colours mismatch")
|
||||
quit()
|
||||
}
|
||||
#=======================================================================
|
||||
#=======
|
||||
# output
|
||||
#=======
|
||||
# plot name and location
|
||||
print(paste0("plot will be in:", plotdir))
|
||||
bp_aa_subcols_ligand = "barplot_acoloured_LIG.svg"
|
||||
plot_bp_aa_subcols_ligand = paste0(plotdir, "/", bp_aa_subcols_ligand)
|
||||
|
||||
#=======================================================================
|
||||
#================
|
||||
# Data for plots
|
||||
#================
|
||||
# REASSIGNMENT as necessary
|
||||
df = my_df_u_cols_lig
|
||||
|
||||
cat("ligand df dim:", dim(df))
|
||||
|
||||
# sanity checks
|
||||
str(df)
|
||||
|
||||
# sanity check
|
||||
df[df$position == "49",]
|
||||
df[df$position == "13",]
|
||||
df[df$position == "103",]
|
||||
|
||||
###########################
|
||||
# Plot: Ligand affinity
|
||||
###########################
|
||||
|
||||
#==========================
|
||||
# Barplot with scores (unordered)
|
||||
# corresponds to ligand_outcome
|
||||
# Stacked Barplot with colours: ligand_outcome @ position coloured by
|
||||
# stability scores. This is a barplot where each bar corresponds
|
||||
# to a SNP and is coloured by its corresponding ligand stability value.
|
||||
# Normalised values (range between -1 and 1 ) to aid visualisation
|
||||
# NOTE: since barplot plots discrete values, colour = score, so number of
|
||||
# colours will be equal to the no. of unique normalised scores
|
||||
# rather than a continuous scale
|
||||
# will require generating the colour scale separately.
|
||||
#============================
|
||||
# sanity checks
|
||||
upos = unique(df$position)
|
||||
|
||||
table(df$ligand_outcome)
|
||||
table(df$ligand_outcome)
|
||||
|
||||
# add frequency of positions (from lib data.table)
|
||||
setDT(df)[, pos_count := .N, by = .(position)]
|
||||
|
||||
# this is cummulative
|
||||
table(df$pos_count)
|
||||
|
||||
# use group by on this
|
||||
library(dplyr)
|
||||
snpsBYpos_df <- df %>%
|
||||
group_by(position) %>%
|
||||
summarize(snpsBYpos = mean(pos_count))
|
||||
|
||||
table(snpsBYpos_df$snpsBYpos)
|
||||
|
||||
snp_count = sort(unique(snpsBYpos_df$snpsBYpos))
|
||||
|
||||
# sanity checks
|
||||
# should be a factor
|
||||
if (is.factor(df$ligand_outcome)){
|
||||
print("ligand_outcome is factor")
|
||||
}else{
|
||||
print("converting ligand_outcome to factor")
|
||||
df$ligand_outcome = as.factor(df$ligand_outcome)
|
||||
}
|
||||
|
||||
is.factor(df$ligand_outcome)
|
||||
|
||||
table(df$ligand_outcome)
|
||||
|
||||
# may not be -1 and 1 since these are filtered within 10A
|
||||
min(df$affinity_scaled)
|
||||
max(df$affinity_scaled)
|
||||
|
||||
# sanity checks
|
||||
# very important!!!!
|
||||
tapply(df$affinity_scaled, df$ligand_outcome, min)
|
||||
|
||||
tapply(df$affinity_scaled, df$ligand_outcome, max)
|
||||
|
||||
# My colour FUNCTION: based on group and subgroup
|
||||
# in my case;
|
||||
# df = df
|
||||
# group = ligand_outcome
|
||||
# subgroup = normalised score i.e affinity_scaled
|
||||
|
||||
# check unique values in normalised data
|
||||
u = unique(df$affinity_scaled)
|
||||
cat("No. of unique values in normalised data:", length(u))
|
||||
|
||||
# Define group
|
||||
# Create an extra column called group which contains the "gp name and score"
|
||||
# so colours can be generated for each unique values in this column
|
||||
my_grp = df$affinity_scaled
|
||||
df$group <- paste0(df$ligand_outcome, "_", my_grp, sep = "")
|
||||
|
||||
# Call the function to create the palette based on the group defined above
|
||||
colours <- ColourPalleteMulti(df, "ligand_outcome", "my_grp")
|
||||
print(paste0("Colour palette generated for: ", length(colours), " colours"))
|
||||
my_title = "Ligand affinity"
|
||||
cat("No. of axis colours: ", length(my_axis_colours))
|
||||
|
||||
#========================
|
||||
# plot with axis colours
|
||||
#========================
|
||||
class(df$lab_bg)
|
||||
|
||||
# define cartesian coord
|
||||
my_xlim = length(unique(df$position)); my_xlim
|
||||
|
||||
# axis label size
|
||||
my_xals = 18
|
||||
my_yals = 18
|
||||
|
||||
# axes text size
|
||||
my_xats = 14
|
||||
my_yats = 18
|
||||
|
||||
#******************
|
||||
# generate plot: with axis colours
|
||||
#******************
|
||||
print(paste0("plot name:", plot_bp_aa_subcols_ligand))
|
||||
|
||||
svg(plot_bp_aa_subcols_ligand, width = 26, height = 4)
|
||||
|
||||
g = ggplot(df, aes(factor(position, ordered = T)))
|
||||
|
||||
outPlot = g +
|
||||
coord_cartesian(xlim = c(1, my_xlim)
|
||||
#, ylim = c(0, 6)
|
||||
, ylim = c(0, max(snp_count))
|
||||
, clip = "off") +
|
||||
geom_bar(aes(fill = group), colour = "grey") +
|
||||
scale_fill_manual(values = colours
|
||||
, guide = "none") +
|
||||
geom_tile(aes(,-0.8, width = 0.95, height = 0.85)
|
||||
, fill = df$lab_bg) +
|
||||
geom_tile(aes(,-1.2, width = 0.95, height = -0.2)
|
||||
, fill = df$lab_bg2) +
|
||||
|
||||
# Here it"s important to specify that your axis goes from 1 to max number of levels
|
||||
theme(axis.text.x = element_text(size = my_xats
|
||||
, angle = 90
|
||||
, hjust = 1
|
||||
, vjust = 0.4
|
||||
, colour = my_axis_colours)
|
||||
|
||||
, axis.text.y = element_text(size = my_yats
|
||||
, angle = 0
|
||||
, hjust = 1
|
||||
, vjust = 0)
|
||||
, axis.title.x = element_text(size = my_xals)
|
||||
#, hjust = 1
|
||||
#, vjust = 0.4)
|
||||
, axis.title.y = element_text(size = my_yals )
|
||||
, axis.ticks.x = element_blank()) +
|
||||
labs(title = ""
|
||||
#title = my_title
|
||||
, x = "position"
|
||||
, y = "Frequency")
|
||||
|
||||
print(outPlot)
|
||||
dev.off()
|
||||
|
||||
#!!!!!!!!!!!!!!!!
|
||||
#Warning message:
|
||||
# Vectorized input to `element_text()` is not officially supported.
|
||||
#Results may be unexpected or may change in future versions of ggplot2.
|
||||
#!!!!!!!!!!!!!!!!!
|
||||
|
||||
# for sanity and good practice
|
||||
#rm(df)
|
||||
Loading…
Add table
Add a link
Reference in a new issue