tanu/LSHTM_analysis
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity
LSHTM_analysis/scripts/plotting/logo_plots.R

276 lines
No EOL
9.1 KiB
R
Executable file
Raw Blame History

#!/usr/bin/env Rscript
#########################################################
# TASK: producing logo-type plot showing
## OR
## Log OR
# multiple muts (>1 per position) coloured by aa property
## mutant_type
## wild_type
# Note: Replaced scripts:
## logo_plot.R
## logo_multiple_muts.R
## logo_combined.R
#########################################################
getwd()
setwd("~/git/LSHTM_analysis/scripts/plotting")
getwd()
source("Header_TT.R")
spec = matrix(c(
"drug" , "d", 1, "character",
"gene" , "g", 1, "character",
"data_file1" , "fa", 2, "character",
"data_file2" , "fb", 2, "character"
), byrow = TRUE, ncol = 4)
opt = getopt(spec)
drug = opt$drug
gene = opt$gene
infile_params = opt$data_file1
infile_metadata = opt$data_file2
if(is.null(drug)|is.null(gene)) {
stop("Missing arguments: --drug and --gene must both be specified (case-sensitive)")
}
#===========
# Input
#===========
source("get_plotting_dfs.R")
#===========
# output
#===========
logo_or_plotname = "logo_or_plot.svg"
plot_logo_or = paste0(plotdir,"/", logo_or_plotname)
logo_logOR_plotname = "logo_logOR_plot.svg"
plot_logo_logOR = paste0(plotdir,"/", logo_logOR_plotname)
logo_multiple_muts = "logo_multiple_muts.svg"
plot_logo_multiple_muts = paste0(plotdir,"/", logo_multiple_muts)
logo_combined_labelled = "logo_combined_labelled.svg"
plot_logo_combined_labelled = paste0(plotdir,"/", logo_combined_labelled)
#########################################################
#logo_or_mult_p + theme_dark()
#logo_or_mult_p + theme(plot.background = element_rect(fill = "black"))
#==================================
# Output
# Logo plot OR: custom height (OR)
#==================================
cat("Logo plot with OR as y axis:", plot_logo_or)
svg(plot_logo_or, width = 30 , height = 6)
logo_or = ggseqlogo(wide_df_or
, method = "custom"
, seq_type = "aa") + ylab("my custom height") +
theme( axis.text.x = element_text(size = 12
, angle = 90
, hjust = 1
, vjust = 0.4)
, axis.text.y = element_text(size = 22
, angle = 0
, hjust = 1
, vjust = 0)
, axis.title.y = element_text(size = 25)
, axis.title.x = element_text(size = 20)
#, legend.position = "bottom") +
, legend.position = "none")+
#, legend.text = element_text(size = 15)
#, legend.title = element_text(size = 15))+
scale_x_discrete("Position"
#, breaks
, labels = position_or
, limits = factor(1:length(position_or))) +
ylab("Odds Ratio")
print(logo_or)
#dev.off()
#=============================================
# Output
# Logo plot LOG OR: custom height (Log10 OR)
#=============================================
cat("Logo plot with log10 OR as y axis:", plot_logo_logOR)
svg(plot_logo_logOR, width = 30 , height = 6)
logo_logOR = ggseqlogo(wide_df_logor_m
, method = "custom"
, seq_type="aa") + ylab("my custom height") +
theme(legend.position = "bottom"
, axis.text.x = element_text(size = 13
, angle = 90
, hjust = 1
, vjust = 0.4)
, axis.text.y = element_text(size = 20
, angle = 0
, hjust = 1
, vjust = 0)
, axis.title.y = element_text(size = 25)
, axis.title.x = element_text(size = 20))+
scale_x_discrete("Position"
#, breaks
, labels = position_logor
, limits = factor(1:length(position_logor)))+
ylab("Log (Odds Ratio)") +
scale_y_continuous(limits = c(0, 9))
print(logo_logOR)
#dev.off()
#*****************************
# Mutant logo plot: >1 nsSNP
#******************************
aa_col_choices = c('chemistry', 'hydrophobicity', 'clustalx', 'taylor')
my_logo_col = aa_col_choices[[1]]
if (my_logo_col == 'clustalx || taylor'){
cat("\nSelected colour scheme:", my_logo_col
, "\nUsing black theme\n")
theme_bgc = "black"
font_bgc = "white"
} if (my_logo_col == 'chemistry || hydrophobicity') {
cat('\nSelected colour scheme:', my_logo_col
, "\nUsing grey theme")
theme_bgc = "grey"
font_bgc = "black"
}
p0 = ggseqlogo(tab_mt
, method = 'custom'
, col_scheme = my_logo_col
, seq_type = 'aa') +
#ylab('my custom height') +
theme(axis.text.x = element_blank()) +
theme(text=element_text(family="FreeSans"))+
theme_logo()+
scale_x_continuous(breaks = 1:ncol(tab_mt)
, labels = colnames(tab_mt))+
scale_y_continuous( breaks = 1:max_mult_mut
, limits = c(0, max_mult_mut))
#p0
cat('\nDone: p0')
# further customisation
mut_logo_p = p0 + theme(legend.position = "none"
, legend.title = element_blank()
, legend.text = element_text(size = 20)
, axis.text.x = element_text(size = 14
, angle = 90
, colour = font_bgc)
, axis.text.y = element_blank()
, plot.background = element_rect(fill = theme_bgc))
mut_logo_p
cat('\nDone: p0+mut_logo_p')
#*************************
# Wild logo plot: >1 nsSNP
#*************************
p2 = ggseqlogo(tab_wt
, method = 'custom'
, seq_type = 'aa'
, col_scheme = my_logo_col) +
#ylab('my custom height') +
theme(text=element_text(family="FreeSans"))+
theme(axis.text.x = element_blank()
, axis.text.y = element_blank()) +
theme_logo() +
scale_x_continuous(breaks = 1:ncol(tab_wt)
, labels = colnames(tab_wt))
#p2
cat('\nDone: p2')
# further customise
wt_logo_p = p2 +
theme(legend.position = "bottom"
#, legend.title = element_blank()
, legend.title = element_text("Amino acid properties", size = 20)
, legend.text = element_text(size = 20)
, axis.text.x = element_text(size = 14, angle = 90)
, axis.text.y = element_blank()
, axis.title.x = element_text(size = 22))+
labs(x= "Position")
#wt_logo_p
cat('\nDone: wt_logo_p')
#***********************
# Logo OR >1 nsSNP
#***********************
logo_or_mult_p = ggseqlogo(wide_df_or_mult
, method = "custom"
, col_scheme = my_logo_col
, seq_type="aa") +
ylab("my custom height") +
theme(axis.text.x = element_text(size = 14
, angle = 90
, hjust = 1
, vjust = 0.4)
, axis.text.y = element_text(size = 18
, angle = 0
, hjust = 1
, vjust = 0)
, axis.title.y = element_text(size = 18)
, axis.title.x = element_blank()
, legend.position = "none")+
scale_x_discrete( labels = position_or_mult
, limits = factor(1:length(position_or_mult))) +
scale_y_discrete(breaks = c(50, 150, 250, 350)
, labels = c(50, 150, 250, 350)
, limits = c(50, 150, 250, 350)
) +
xlab("Position") +
ylab("Odds Ratio")
#logo_or_mult_p
cat('\nDone: logo_or_mult_p')
#=========================================
# Output
# Combined plot: logo_mutliple_muts.svg
#=========================================
#suppressMessages( require(cowplot) )
cat('\nDone: wt_logo_p')
#plot_grid(p1, p3, ncol = 1, align = 'v')
cat('\nDone: mut_logo_p + wt_logo_p')
# colour scheme: https://rdrr.io/cran/ggseqlogo/src/R/col_schemes.r
cat("\nOutput plot:", plot_logo_multiple_muts, "\n")
svg(plot_logo_multiple_muts, width = 32, height = 10)
mult_muts_combined = cowplot::plot_grid(mut_logo_p, wt_logo_p
, nrow = 2
, align = "v"
, rel_heights = c(3/4, 1/4))
print(mult_muts_combined)
#dev.off()
#======================================================
# Output
# Combined ALL logo plots: logOR, mut_logo and mut_logo
#=====================================================
cat("Output plot:", plot_logo_combined_labelled)
svg(plot_logo_combined_labelled, width = 25, height = 10)
all_logo_plots = cowplot::plot_grid(logo_or_mult_p
, mut_logo_p
, wt_logo_p
, nrow = 3
, align = "hv"
#, labels = c("(a)","(b)", "(c)")
, labels = "AUTO"
, rel_heights = c(3/8, 3/8, 1.5/8)
, rel_widths = c(0.85, 1, 1)
, label_size = 25)
print(all_logo_plots)
#dev.off()
Reference in a new issue View git blame Copy permalink