470 lines
17 KiB
R
Executable file
470 lines
17 KiB
R
Executable file
#!/usr/bin/env Rscript
|
|
#########################################################
|
|
# TASK: Script to format data for dm om plots:
|
|
# generating LF data
|
|
# sourced by get_plotting_dfs.R
|
|
#########################################################
|
|
# working dir and loading libraries
|
|
# getwd()
|
|
# setwd("~/git/LSHTM_analysis/scripts/plotting")
|
|
# getwd()
|
|
|
|
# make cmd
|
|
# globals
|
|
# drug = "streptomycin"
|
|
# gene = "gid"
|
|
|
|
# source("get_plotting_dfs.R")
|
|
#=======================================================================
|
|
# MOVE TO COMBINE or singular file for deepddg
|
|
#
|
|
# cols_to_select = c("mutation", "mutationinformation"
|
|
# , "wild_type", "position", "mutant_type"
|
|
# , "mutation_info")
|
|
#
|
|
# merged_df3_short = merged_df3[, cols_to_select]
|
|
|
|
# infilename_mcsm_f_snps <- paste0("~/git/Data/", drug, "/output/", gene
|
|
# , "_mcsm_formatted_snps.csv")
|
|
#
|
|
# mcsm_f_snps<- read.csv(infilename_mcsm_f_snps, header = F)
|
|
# names(mcsm_f_snps) <- "mutationinformation"
|
|
|
|
# write merged_df3 to generate structural figure on chimera
|
|
#write.csv(merged_df3_short, "merged_df3_short.csv")
|
|
#========================================================================
|
|
|
|
#========================================================================
|
|
# cols to select
|
|
|
|
cols_mcsm_df <- merged_df3[, c("mutationinformation", "mutation"
|
|
, "mutation_info", "position"
|
|
, LigDist_colname
|
|
, "duet_stability_change", "duet_scaled", "duet_outcome"
|
|
, "ligand_affinity_change", "affinity_scaled", "ligand_outcome"
|
|
, "ddg_foldx", "foldx_scaled", "foldx_outcome"
|
|
, "deepddg", "deepddg_scaled", "deepddg_outcome"
|
|
, "asa", "rsa"
|
|
, "rd_values", "kd_values"
|
|
, "log10_or_mychisq", "neglog_pval_fisher", "af")]
|
|
|
|
cols_mcsm_na_df <- mcsm_na_df[, c("mutationinformation"
|
|
, "mcsm_na_affinity", "mcsm_na_scaled"
|
|
, "mcsm_na_outcome")]
|
|
# entire dynamut_df
|
|
|
|
cols_dynamut2_df <- dynamut2_df[, c("mutationinformation"
|
|
, "ddg_dynamut2", "ddg_dynamut2_scaled"
|
|
, "ddg_dynamut2_outcome")]
|
|
|
|
n_comb_cols = length(cols_mcsm_df) + length(cols_mcsm_na_df) +
|
|
length(dynamut_df) + length(cols_dynamut2_df); n_comb_cols
|
|
|
|
i1<- intersect(names(cols_mcsm_df), names(cols_mcsm_na_df))
|
|
i2<- intersect(names(dynamut_df), names(cols_dynamut2_df))
|
|
merging_cols <- intersect(i1, i2)
|
|
cat("\nmerging_cols:", merging_cols)
|
|
|
|
if (merging_cols == "mutationinformation") {
|
|
cat("\nStage 1: Found common col between dfs, checking values in it...")
|
|
c1 <- all(mcsm_f_snps[[merging_cols]]%in%cols_mcsm_df[[merging_cols]])
|
|
c2 <- all(mcsm_f_snps[[merging_cols]]%in%cols_mcsm_na_df[[merging_cols]])
|
|
c3 <- all(mcsm_f_snps[[merging_cols]]%in%dynamut_df[[merging_cols]])
|
|
c4 <- all(mcsm_f_snps[[merging_cols]]%in%cols_dynamut2_df[[merging_cols]])
|
|
cols_check <- c(c1, c2, c3, c4)
|
|
expected_cols = n_comb_cols - ( length(cols_check) - 1)
|
|
if (all(cols_check)){
|
|
cat("\nStage 2: Proceeding with merging dfs:\n")
|
|
comb_df <- Reduce(inner_join, list(cols_mcsm_df
|
|
, cols_mcsm_na_df
|
|
, dynamut_df
|
|
, cols_dynamut2_df))
|
|
comb_df_s = arrange(comb_df, position)
|
|
|
|
# if ( nrow(comb_df_s) == nrow(mcsm_f_snps) && ncol(comb_df_s) == expected_cols) {
|
|
# cat("\Stage3, PASS: dfs merged sucessfully"
|
|
# , "\nnrow of merged_df: ", nrow(comb_df_s)
|
|
# , "\nncol of merged_df:", ncol(comb_df_s))
|
|
# }
|
|
|
|
}
|
|
}
|
|
#names(comb_df_s)
|
|
cat("\n!!!IT GOT TO HERE!!!!")
|
|
#=======================================================================
|
|
fact_cols = colnames(comb_df_s)[grepl( "_outcome|_info", colnames(comb_df_s) )]
|
|
fact_cols
|
|
lapply(comb_df_s[, fact_cols], class)
|
|
comb_df_s[, fact_cols] <- lapply(comb_df_s[, fact_cols], as.factor)
|
|
|
|
if (any(lapply(comb_df_s[, fact_cols], class) == "character")){
|
|
cat("\nChanging cols to factor")
|
|
comb_df_s[, fact_cols] <- lapply(comb_df_s[, fact_cols],as.factor)
|
|
if (all(lapply(comb_df_s[, fact_cols], class) == "factor")){
|
|
cat("\nSuccessful: cols changed to factor")
|
|
}
|
|
}
|
|
lapply(comb_df_s[, fact_cols], class)
|
|
|
|
#=======================================================================
|
|
table(comb_df_s$mutation_info)
|
|
|
|
# further checks to make sure dr and other muts are indeed unique
|
|
dr_muts = comb_df_s[comb_df_s$mutation_info == dr_muts_col,]
|
|
dr_muts_names = unique(dr_muts$mutation)
|
|
|
|
other_muts = comb_df_s[comb_df_s$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 and others muts are NOT unique!")
|
|
quit()
|
|
}
|
|
|
|
# pretty display names i.e. labels to reduce major code duplication later
|
|
foo_cnames = data.frame(colnames(comb_df_s))
|
|
names(foo_cnames) <- "old_name"
|
|
|
|
stability_suffix <- paste0(delta_symbol, delta_symbol, "G")
|
|
flexibility_suffix <- paste0(delta_symbol, delta_symbol, "S")
|
|
|
|
lig_dn = paste0("Ligand distance (", angstroms_symbol, ")"); lig_dn
|
|
duet_dn = paste0("DUET ", stability_suffix); duet_dn
|
|
foldx_dn = paste0("FoldX ", stability_suffix); foldx_dn
|
|
deepddg_dn = paste0("Deepddg " , stability_suffix); deepddg_dn
|
|
mcsm_na_dn = paste0("mCSM-NA affinity ", stability_suffix); mcsm_na_dn
|
|
dynamut_dn = paste0("Dynamut ", stability_suffix); dynamut_dn
|
|
dynamut2_dn = paste0("Dynamut2 " , stability_suffix); dynamut2_dn
|
|
encom_ddg_dn = paste0("EnCOM " , stability_suffix); encom_ddg_dn
|
|
encom_dds_dn = paste0("EnCOM " , flexibility_suffix ); encom_dds_dn
|
|
sdm_dn = paste0("SDM " , stability_suffix); sdm_dn
|
|
mcsm_dn = paste0("mCSM " , stability_suffix ); mcsm_dn
|
|
|
|
# Change colnames of some columns using datatable
|
|
comb_df_sl = comb_df_s
|
|
names(comb_df_sl)
|
|
|
|
setnames(comb_df_sl
|
|
, old = c("asa", "rsa", "rd_values", "kd_values"
|
|
, "log10_or_mychisq", "neglog_pval_fisher", "af"
|
|
, LigDist_colname
|
|
, "duet_scaled"
|
|
, "foldx_scaled"
|
|
, "deepddg_scaled"
|
|
, "mcsm_na_scaled"
|
|
, "ddg_dynamut_scaled"
|
|
, "ddg_dynamut2_scaled"
|
|
, "ddg_encom_scaled"
|
|
, "dds_encom_scaled"
|
|
, "ddg_sdm"
|
|
, "ddg_mcsm")
|
|
|
|
, new = c("ASA", "RSA", "RD", "KD"
|
|
, "Log10 (OR)", "-Log (P)", "MAF"
|
|
, lig_dn
|
|
, duet_dn
|
|
, foldx_dn
|
|
, deepddg_dn
|
|
, mcsm_na_dn
|
|
, dynamut_dn
|
|
, dynamut2_dn
|
|
, encom_ddg_dn
|
|
, encom_dds_dn
|
|
, sdm_dn
|
|
, mcsm_dn)
|
|
)
|
|
|
|
foo_cnames <- cbind(foo_cnames, colnames(comb_df_sl))
|
|
|
|
# some more pretty labels
|
|
table(comb_df_sl$mutation_info)
|
|
|
|
levels(comb_df_sl$mutation_info)[levels(comb_df_sl$mutation_info)==dr_muts_col] <- "DM"
|
|
levels(comb_df_sl$mutation_info)[levels(comb_df_sl$mutation_info)==other_muts_col] <- "OM"
|
|
|
|
table(comb_df_sl$mutation_info)
|
|
|
|
#######################################################################
|
|
#======================
|
|
# Selecting dfs
|
|
# with appropriate cols
|
|
#=======================
|
|
static_cols_start = c("mutationinformation"
|
|
, "position"
|
|
, "mutation"
|
|
, "mutation_info")
|
|
|
|
static_cols_end = c(lig_dn
|
|
, "ASA"
|
|
, "RSA"
|
|
, "RD"
|
|
, "KD")
|
|
|
|
# ordering is important!
|
|
|
|
#########################################################################
|
|
#==============
|
|
# DUET: LF
|
|
#==============
|
|
cols_to_select_duet = c(static_cols_start, c("duet_outcome", duet_dn), static_cols_end)
|
|
wf_duet = comb_df_sl[, cols_to_select_duet]
|
|
|
|
#pivot_cols_ps = cols_to_select_ps[1:5]; pivot_cols_ps
|
|
pivot_cols_duet = cols_to_select_duet[1: (length(static_cols_start) + 1)]; pivot_cols_duet
|
|
|
|
expected_rows_lf = nrow(wf_duet) * (length(wf_duet) - length(pivot_cols_duet))
|
|
expected_rows_lf
|
|
|
|
# LF data: duet
|
|
lf_duet = gather(wf_duet
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(duet_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_duet) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for ", duet_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
|
|
############################################################################
|
|
#==============
|
|
# FoldX: LF
|
|
#==============
|
|
cols_to_select_foldx= c(static_cols_start, c("foldx_outcome", foldx_dn), static_cols_end)
|
|
wf_foldx = comb_df_sl[, cols_to_select_foldx]
|
|
|
|
pivot_cols_foldx = cols_to_select_foldx[1: (length(static_cols_start) + 1)]; pivot_cols_foldx
|
|
|
|
expected_rows_lf = nrow(wf_foldx) * (length(wf_foldx) - length(pivot_cols_foldx))
|
|
expected_rows_lf
|
|
|
|
# LF data: duet
|
|
print("TESTXXXXXXXXXXXXXXXXXXXXX---------------------->>>>")
|
|
lf_foldx <<- gather(wf_foldx
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(foldx_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_foldx) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for ", foldx_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
|
|
############################################################################
|
|
#==============
|
|
# Deepddg: LF
|
|
#==============
|
|
cols_to_select_deepddg = c(static_cols_start, c("deepddg_outcome", deepddg_dn), static_cols_end)
|
|
wf_deepddg = comb_df_sl[, cols_to_select_deepddg]
|
|
|
|
pivot_cols_deepddg = cols_to_select_deepddg[1: (length(static_cols_start) + 1)]; pivot_cols_deepddg
|
|
|
|
expected_rows_lf = nrow(wf_deepddg) * (length(wf_deepddg) - length(pivot_cols_deepddg))
|
|
expected_rows_lf
|
|
|
|
# LF data: duet
|
|
lf_deepddg = gather(wf_deepddg
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(deepddg_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_deepddg) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for ", deepddg_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
|
|
############################################################################
|
|
#==============
|
|
# mCSM-NA: LF
|
|
#==============
|
|
cols_to_select_mcsm_na = c(static_cols_start, c("mcsm_na_outcome", mcsm_na_dn), static_cols_end)
|
|
wf_mcsm_na = comb_df_sl[, cols_to_select_mcsm_na]
|
|
|
|
pivot_cols_mcsm_na = cols_to_select_mcsm_na[1: (length(static_cols_start) + 1)]; pivot_cols_mcsm_na
|
|
|
|
expected_rows_lf = nrow(wf_mcsm_na) * (length(wf_mcsm_na) - length(pivot_cols_mcsm_na))
|
|
expected_rows_lf
|
|
|
|
# LF data: duet
|
|
lf_mcsm_na = gather(wf_mcsm_na
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(mcsm_na_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_mcsm_na) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for ", mcsm_na_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
|
|
############################################################################
|
|
#==============
|
|
# Dynamut: LF
|
|
#==============
|
|
cols_to_select_dynamut = c(static_cols_start, c("ddg_dynamut_outcome", dynamut_dn), static_cols_end)
|
|
wf_dynamut = comb_df_sl[, cols_to_select_dynamut]
|
|
|
|
pivot_cols_dynamut = cols_to_select_dynamut[1: (length(static_cols_start) + 1)]; pivot_cols_dynamut
|
|
|
|
expected_rows_lf = nrow(wf_dynamut) * (length(wf_dynamut) - length(pivot_cols_dynamut))
|
|
expected_rows_lf
|
|
|
|
# LF data: duet
|
|
lf_dynamut = gather(wf_dynamut
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(dynamut_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_dynamut) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for ", dynamut_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
|
|
############################################################################
|
|
#==============
|
|
# Dynamut2: LF
|
|
#==============
|
|
cols_to_select_dynamut2 = c(static_cols_start, c("ddg_dynamut2_outcome", dynamut2_dn), static_cols_end)
|
|
|
|
wf_dynamut2 = comb_df_sl[, cols_to_select_dynamut2]
|
|
|
|
pivot_cols_dynamut2 = cols_to_select_dynamut2[1: (length(static_cols_start) + 1)]; pivot_cols_dynamut2
|
|
|
|
expected_rows_lf = nrow(wf_dynamut2) * (length(wf_dynamut2) - length(pivot_cols_dynamut2))
|
|
expected_rows_lf
|
|
|
|
# LF data: duet
|
|
lf_dynamut2 = gather(wf_dynamut2
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(dynamut2_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_dynamut2) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for ", dynamut2_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
|
|
############################################################################
|
|
#==============
|
|
# EnCOM ddg: LF
|
|
#==============
|
|
cols_to_select_encomddg = c(static_cols_start, c("ddg_encom_outcome", encom_ddg_dn), static_cols_end)
|
|
wf_encomddg = comb_df_sl[, cols_to_select_encomddg]
|
|
|
|
pivot_cols_encomddg = cols_to_select_encomddg[1: (length(static_cols_start) + 1)]; pivot_cols_encomddg
|
|
|
|
expected_rows_lf = nrow(wf_encomddg ) * (length(wf_encomddg ) - length(pivot_cols_encomddg))
|
|
expected_rows_lf
|
|
|
|
# LF data: encomddg
|
|
lf_encomddg = gather(wf_encomddg
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(encom_ddg_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_encomddg) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for ", encom_ddg_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
############################################################################
|
|
#==============
|
|
# EnCOM dds: LF
|
|
#==============
|
|
cols_to_select_encomdds = c(static_cols_start, c("dds_encom_outcome", encom_dds_dn), static_cols_end)
|
|
wf_encomdds = comb_df_sl[, cols_to_select_encomdds]
|
|
|
|
pivot_cols_encomdds = cols_to_select_encomdds[1: (length(static_cols_start) + 1)]; pivot_cols_encomdds
|
|
|
|
expected_rows_lf = nrow(wf_encomdds) * (length(wf_encomdds) - length(pivot_cols_encomdds))
|
|
expected_rows_lf
|
|
|
|
# LF data: encomddg
|
|
lf_encomdds = gather(wf_encomdds
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(encom_dds_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_encomdds) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for", encom_dds_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
|
|
############################################################################
|
|
#==============
|
|
# SDM: LF
|
|
#==============
|
|
cols_to_select_sdm = c(static_cols_start, c("ddg_sdm_outcome", sdm_dn), static_cols_end)
|
|
wf_sdm = comb_df_sl[, cols_to_select_sdm]
|
|
|
|
pivot_cols_sdm = cols_to_select_sdm[1: (length(static_cols_start) + 1)]; pivot_cols_sdm
|
|
|
|
expected_rows_lf = nrow(wf_sdm) * (length(wf_sdm) - length(pivot_cols_sdm))
|
|
expected_rows_lf
|
|
|
|
# LF data: encomddg
|
|
lf_sdm = gather(wf_sdm
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(sdm_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_sdm) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for", sdm_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
|
|
############################################################################
|
|
#==============
|
|
# mCSM: LF
|
|
#==============
|
|
cols_to_select_mcsm = c(static_cols_start, c("ddg_mcsm_outcome", mcsm_dn), static_cols_end)
|
|
wf_mcsm = comb_df_sl[, cols_to_select_mcsm]
|
|
|
|
pivot_cols_mcsm = cols_to_select_mcsm[1: (length(static_cols_start) + 1)]; pivot_cols_mcsm
|
|
|
|
expected_rows_lf = nrow(wf_mcsm) * (length(wf_mcsm) - length(pivot_cols_mcsm))
|
|
expected_rows_lf
|
|
|
|
# LF data: encomddg
|
|
lf_mcsm = gather(wf_mcsm
|
|
, key = param_type
|
|
, value = param_value
|
|
, all_of(mcsm_dn):tail(static_cols_end,1)
|
|
, factor_key = TRUE)
|
|
|
|
if (nrow(lf_mcsm) == expected_rows_lf){
|
|
cat("\nPASS: long format data created for", mcsm_dn)
|
|
}else{
|
|
cat("\nFAIL: long format data could not be created for duet")
|
|
quit()
|
|
}
|
|
############################################################################
|
|
|