301 lines
10 KiB
R
301 lines
10 KiB
R
#!/usr/bin/env Rscript
|
|
#########################################################
|
|
# TASK: producing boxplots for dr and other muts
|
|
|
|
#########################################################
|
|
#=======================================================================
|
|
# working dir and loading libraries
|
|
getwd()
|
|
#setwd("~/git/LSHTM_analysis/scripts/plotting")
|
|
getwd()
|
|
|
|
#source("~/git/LSHTM_analysis/scripts/Header_TT.R")
|
|
library(ggplot2)
|
|
library(data.table)
|
|
library(dplyr)
|
|
library(tidyverse)
|
|
#source("combining_dfs_plotting.R")
|
|
|
|
#rm(merged_df2, merged_df2_comp, merged_df2_lig, merged_df2_comp_lig
|
|
# , merged_df3_comp, merged_df3_comp_lig
|
|
# , my_df_u, my_df_u_lig)
|
|
|
|
|
|
cols_to_select = c("mutation", "mutationinformation"
|
|
, "wild_type", "position", "mutant_type"
|
|
, "mutation_info")
|
|
|
|
merged_df3_short = merged_df3[, cols_to_select]
|
|
|
|
# write merged_df3 to generate structural figure
|
|
write.csv(merged_df3_short, "merged_df3_short.csv")
|
|
|
|
#========================================================================
|
|
#%%%%%%%%%%%%%%%%%%%
|
|
# REASSIGNMENT: PS
|
|
#%%%%%%%%%%%%%%%%%%%%
|
|
df_ps = merged_df3
|
|
|
|
#============================
|
|
# adding foldx scaled values
|
|
# scale data b/w -1 and 1
|
|
#============================
|
|
n = which(colnames(df_ps) == "ddg"); n
|
|
|
|
my_min = min(df_ps[,n]); my_min
|
|
my_max = max(df_ps[,n]); my_max
|
|
|
|
df_ps$foldx_scaled = ifelse(df_ps[,n] < 0
|
|
, df_ps[,n]/abs(my_min)
|
|
, df_ps[,n]/my_max)
|
|
# sanity check
|
|
my_min = min(df_ps$foldx_scaled); my_min
|
|
my_max = max(df_ps$foldx_scaled); my_max
|
|
|
|
if (my_min == -1 && my_max == 1){
|
|
cat("PASS: foldx ddg successfully scaled b/w -1 and 1"
|
|
, "\nProceeding with assigning foldx outcome category")
|
|
}else{
|
|
cat("FAIL: could not scale foldx ddg values"
|
|
, "Aborting!")
|
|
}
|
|
|
|
#================================
|
|
# adding foldx outcome category
|
|
# ddg<0 = "Stabilising" (-ve)
|
|
#=================================
|
|
|
|
c1 = table(df_ps$ddg < 0)
|
|
df_ps$foldx_outcome = ifelse(df_ps$ddg < 0, "Stabilising", "Destabilising")
|
|
c2 = table(df_ps$ddg < 0)
|
|
|
|
if ( all(c1 == c2) ){
|
|
cat("PASS: foldx outcome successfully created")
|
|
}else{
|
|
cat("FAIL: foldx outcome could not be created. Aborting!")
|
|
exit()
|
|
}
|
|
#=======================================================================
|
|
# name tidying
|
|
df_ps$mutation_info = as.factor(df_ps$mutation_info)
|
|
df_ps$duet_outcome = as.factor(df_ps$duet_outcome)
|
|
df_ps$foldx_outcome = as.factor(df_ps$foldx_outcome)
|
|
df_ps$ligand_outcome = as.factor(df_ps$ligand_outcome)
|
|
|
|
# check
|
|
table(df_ps$mutation_info)
|
|
|
|
# further checks to make sure dr and other muts are indeed unique
|
|
dr_muts = df_ps[df_ps$mutation_info == dr_muts_col,]
|
|
dr_muts_names = unique(dr_muts$mutation)
|
|
|
|
other_muts = df_ps[df_ps$mutation_info == other_muts_col,]
|
|
other_muts_names = unique(other_muts$mutation)
|
|
|
|
if ( table(dr_muts_names%in%other_muts_names)[[1]] == length(dr_muts_names) &&
|
|
table(other_muts_names%in%dr_muts_names)[[1]] == length(other_muts_names) ){
|
|
cat("PASS: dr and other muts are indeed unique")
|
|
}else{
|
|
cat("FAIL: dr adn others muts are NOT unique!")
|
|
quit()
|
|
}
|
|
|
|
|
|
#%%%%%%%%%%%%%%%%%%%
|
|
# REASSIGNMENT: LIG
|
|
#%%%%%%%%%%%%%%%%%%%%
|
|
|
|
df_lig = merged_df3_lig
|
|
|
|
# name tidying
|
|
df_lig$mutation_info = as.factor(df_lig$mutation_info)
|
|
df_lig$duet_outcome = as.factor(df_lig$duet_outcome)
|
|
#df_lig$ligand_outcome = as.factor(df_lig$ligand_outcome)
|
|
|
|
# check
|
|
table(df_lig$mutation_info)
|
|
|
|
#========================================================================
|
|
#===========
|
|
# Data: ps
|
|
#===========
|
|
# keep similar dtypes cols together
|
|
cols_to_select_ps = c("mutationinformation", "mutation", "position", "mutation_info"
|
|
, "duet_outcome"
|
|
|
|
, "duet_scaled"
|
|
, "ligand_distance"
|
|
, "asa"
|
|
, "rsa"
|
|
, "rd_values"
|
|
, "kd_values")
|
|
|
|
df_wf_ps = df_ps[, cols_to_select_ps]
|
|
|
|
pivot_cols_ps = cols_to_select_ps[1:5]; pivot_cols_ps
|
|
|
|
expected_rows_lf_ps = nrow(df_wf_ps) * (length(df_wf_ps) - length(pivot_cols_ps))
|
|
expected_rows_lf_ps
|
|
|
|
# LF data: duet
|
|
df_lf_ps = gather(df_wf_ps, param_type, param_value, duet_scaled:kd_values, factor_key=TRUE)
|
|
|
|
if (nrow(df_lf_ps) == expected_rows_lf_ps){
|
|
cat("PASS: long format data created for duet")
|
|
}else{
|
|
cat("FAIL: long format data could not be created for duet")
|
|
exit()
|
|
}
|
|
|
|
str(df_wf_ps)
|
|
str(df_lf_ps)
|
|
|
|
# assign pretty labels: param_type
|
|
levels(df_lf_ps$param_type); table(df_lf_ps$param_type)
|
|
|
|
ligand_dist_colname = paste0("Distance to ligand (", angstroms_symbol, ")")
|
|
ligand_dist_colname
|
|
|
|
duet_stability_name = paste0(delta_symbol, delta_symbol, "G")
|
|
duet_stability_name
|
|
|
|
#levels(df_lf_ps$param_type)[levels(df_lf_ps$param_type)=="duet_scaled"] <- "Stability"
|
|
levels(df_lf_ps$param_type)[levels(df_lf_ps$param_type)=="duet_scaled"] <- duet_stability_name
|
|
#levels(df_lf_ps$param_type)[levels(df_lf_ps$param_type)=="ligand_distance"] <- "Ligand Distance"
|
|
levels(df_lf_ps$param_type)[levels(df_lf_ps$param_type)=="ligand_distance"] <- ligand_dist_colname
|
|
levels(df_lf_ps$param_type)[levels(df_lf_ps$param_type)=="asa"] <- "ASA"
|
|
levels(df_lf_ps$param_type)[levels(df_lf_ps$param_type)=="rsa"] <- "RSA"
|
|
levels(df_lf_ps$param_type)[levels(df_lf_ps$param_type)=="rd_values"] <- "RD"
|
|
levels(df_lf_ps$param_type)[levels(df_lf_ps$param_type)=="kd_values"] <- "KD"
|
|
# check
|
|
levels(df_lf_ps$param_type); table(df_lf_ps$param_type)
|
|
|
|
# assign pretty labels: mutation_info
|
|
levels(df_lf_ps$mutation_info); table(df_lf_ps$mutation_info)
|
|
sum(table(df_lf_ps$mutation_info)) == nrow(df_lf_ps)
|
|
|
|
levels(df_lf_ps$mutation_info)[levels(df_lf_ps$mutation_info)==dr_muts_col] <- "DM"
|
|
levels(df_lf_ps$mutation_info)[levels(df_lf_ps$mutation_info)==other_muts_col] <- "OM"
|
|
# check
|
|
levels(df_lf_ps$mutation_info); table(df_lf_ps$mutation_info)
|
|
|
|
############################################################################
|
|
|
|
#===========
|
|
# LF data: LIG
|
|
#===========
|
|
# keep similar dtypes cols together
|
|
cols_to_select_lig = c("mutationinformation", "mutation", "position", "mutation_info"
|
|
, "ligand_outcome"
|
|
|
|
, "affinity_scaled"
|
|
#, "ligand_distance"
|
|
, "asa"
|
|
, "rsa"
|
|
, "rd_values"
|
|
, "kd_values")
|
|
|
|
df_wf_lig = df_lig[, cols_to_select_lig]
|
|
|
|
pivot_cols_lig = cols_to_select_lig[1:5]; pivot_cols_lig
|
|
|
|
expected_rows_lf_lig = nrow(df_wf_lig) * (length(df_wf_lig) - length(pivot_cols_lig))
|
|
expected_rows_lf_lig
|
|
|
|
# LF data: foldx
|
|
df_lf_lig = gather(df_wf_lig, param_type, param_value, affinity_scaled:kd_values, factor_key=TRUE)
|
|
|
|
if (nrow(df_lf_lig) == expected_rows_lf_lig){
|
|
cat("PASS: long format data created for foldx")
|
|
}else{
|
|
cat("FAIL: long format data could not be created for foldx")
|
|
exit()
|
|
}
|
|
|
|
# assign pretty labels: param_type
|
|
levels(df_lf_lig$param_type); table(df_lf_lig$param_type)
|
|
|
|
levels(df_lf_lig$param_type)[levels(df_lf_lig$param_type)=="affinity_scaled"] <- "Ligand Affinity"
|
|
#levels(df_lf_lig$param_type)[levels(df_lf_lig$param_type)=="ligand_distance"] <- "Ligand Distance"
|
|
levels(df_lf_lig$param_type)[levels(df_lf_lig$param_type)=="asa"] <- "ASA"
|
|
levels(df_lf_lig$param_type)[levels(df_lf_lig$param_type)=="rsa"] <- "RSA"
|
|
levels(df_lf_lig$param_type)[levels(df_lf_lig$param_type)=="rd_values"] <- "RD"
|
|
levels(df_lf_lig$param_type)[levels(df_lf_lig$param_type)=="kd_values"] <- "KD"
|
|
#check
|
|
levels(df_lf_lig$param_type); table(df_lf_lig$param_type)
|
|
|
|
# assign pretty labels: mutation_info
|
|
levels(df_lf_lig$mutation_info); table(df_lf_lig$mutation_info)
|
|
sum(table(df_lf_lig$mutation_info)) == nrow(df_lf_lig)
|
|
|
|
levels(df_lf_lig$mutation_info)[levels(df_lf_lig$mutation_info)==dr_muts_col] <- "DM"
|
|
levels(df_lf_lig$mutation_info)[levels(df_lf_lig$mutation_info)==other_muts_col] <- "OM"
|
|
# check
|
|
levels(df_lf_lig$mutation_info); table(df_lf_lig$mutation_info)
|
|
|
|
#############################################################################
|
|
#===========
|
|
# Data: foldx
|
|
#===========
|
|
# keep similar dtypes cols together
|
|
cols_to_select_foldx = c("mutationinformation", "mutation", "position", "mutation_info"
|
|
, "foldx_outcome"
|
|
|
|
, "foldx_scaled")
|
|
#, "ligand_distance"
|
|
#, "asa"
|
|
#, "rsa"
|
|
#, "rd_values"
|
|
#, "kd_values")
|
|
|
|
|
|
df_wf_foldx = df_ps[, cols_to_select_foldx]
|
|
|
|
pivot_cols_foldx = cols_to_select_foldx[1:5]; pivot_cols_foldx
|
|
|
|
expected_rows_lf_foldx = nrow(df_wf_foldx) * (length(df_wf_foldx) - length(pivot_cols_foldx))
|
|
expected_rows_lf_foldx
|
|
|
|
# LF data: foldx
|
|
df_lf_foldx = gather(df_wf_foldx, param_type, param_value, foldx_scaled, factor_key=TRUE)
|
|
|
|
if (nrow(df_lf_foldx) == expected_rows_lf_foldx){
|
|
cat("PASS: long format data created for foldx")
|
|
}else{
|
|
cat("FAIL: long format data could not be created for foldx")
|
|
exit()
|
|
}
|
|
|
|
foldx_stability_name = paste0(delta_symbol, delta_symbol, "G")
|
|
foldx_stability_name
|
|
|
|
# assign pretty labels: param type
|
|
levels(df_lf_foldx$param_type); table(df_lf_foldx$param_type)
|
|
|
|
#levels(df_lf_foldx$param_type)[levels(df_lf_foldx$param_type)=="foldx_scaled"] <- "Stability"
|
|
levels(df_lf_foldx$param_type)[levels(df_lf_foldx$param_type)=="foldx_scaled"] <- foldx_stability_name
|
|
#levels(df_lf_foldx$param_type)[levels(df_lf_foldx$param_type)=="ligand_distance"] <- "Ligand Distance"
|
|
#levels(df_lf_foldx$param_type)[levels(df_lf_foldx$param_type)=="asa"] <- "ASA"
|
|
#levels(df_lf_foldx$param_type)[levels(df_lf_foldx$param_type)=="rsa"] <- "RSA"
|
|
#levels(df_lf_foldx$param_type)[levels(df_lf_foldx$param_type)=="rd_values"] <- "RD"
|
|
#levels(df_lf_foldx$param_type)[levels(df_lf_foldx$param_type)=="kd_values"] <- "KD"
|
|
# check
|
|
levels(df_lf_foldx$param_type); table(df_lf_foldx$param_type)
|
|
|
|
# assign pretty labels: mutation_info
|
|
levels(df_lf_foldx$mutation_info); table(df_lf_foldx$mutation_info)
|
|
sum(table(df_lf_foldx$mutation_info)) == nrow(df_lf_foldx)
|
|
|
|
levels(df_lf_foldx$mutation_info)[levels(df_lf_foldx$mutation_info)==dr_muts_col] <- "DM"
|
|
levels(df_lf_foldx$mutation_info)[levels(df_lf_foldx$mutation_info)==other_muts_col] <- "OM"
|
|
# check
|
|
levels(df_lf_foldx$mutation_info); table(df_lf_foldx$mutation_info)
|
|
|
|
############################################################################
|
|
|
|
# clear excess variables
|
|
rm(cols_to_select_ps, cols_to_select_foldx, cols_to_select_lig
|
|
, pivot_cols_ps, pivot_cols_foldx, pivot_cols_lig
|
|
, expected_rows_lf_ps, expected_rows_lf_foldx, expected_rows_lf_lig
|
|
, my_max, my_min, na_count, na_count_df2, na_count_df3, dup_muts_nu
|
|
, c1, c2, n)
|