From 6d9412d23266ed1833629c84133e8b7b8c0eafd5 Mon Sep 17 00:00:00 2001 From: Tanushree Tunstall Date: Thu, 26 Aug 2021 16:35:46 +0100 Subject: [PATCH] playing with dm_om (other)plots data and graph on gid branch --- scripts/functions/plotting_globals.R | 3 +- scripts/plotting/get_plotting_dfs.R | 203 +------ scripts/plotting/other_plots_combined.R | 13 +- scripts/plotting/other_plots_data.R | 693 ++++++++++++++++-------- 4 files changed, 502 insertions(+), 410 deletions(-) diff --git a/scripts/functions/plotting_globals.R b/scripts/functions/plotting_globals.R index cfd2848..c28047e 100644 --- a/scripts/functions/plotting_globals.R +++ b/scripts/functions/plotting_globals.R @@ -32,7 +32,8 @@ import_dirs <- function(drug_name, gene_name) { #=============================== # mcsm ligand distance cut off #=============================== -#mcsm_lig_cutoff <<- 10 +LigDist_colname <<- "ligand_distance" +LigDist_cutoff <<- 10 #================== # Angstroms symbol diff --git a/scripts/plotting/get_plotting_dfs.R b/scripts/plotting/get_plotting_dfs.R index a9e78e9..2dae471 100644 --- a/scripts/plotting/get_plotting_dfs.R +++ b/scripts/plotting/get_plotting_dfs.R @@ -25,8 +25,8 @@ source("../functions/bp_subcolours.R") # variables for lig #==================== -LigDist_colname = "ligand_distance" -LigDist_cutoff = 10 +#LigDist_colname = "ligand_distance" +#LigDist_cutoff = 10 #=========== # input @@ -54,10 +54,15 @@ pd_df = plotting_data(mcsm_df , lig_dist_colname = LigDist_colname , lig_dist_cutoff = LigDist_cutoff) -my_df = pd_df[[1]] -my_df_u = pd_df[[2]] # this forms one of the input for combining_dfs_plotting() -my_df_u_lig = pd_df[[3]] -dup_muts = pd_df[[4]] +my_df = pd_df[[1]] +my_df_u = pd_df[[2]] # this forms one of the input for combining_dfs_plotting() + +max_ang <- round(max(my_df_u[LigDist_colname])) +min_ang <- round(min(my_df_u[LigDist_colname])) + +cat("\nLigand distance cut off, colname:", LigDist_colname + , "\nThe max distance", gene, "structure df" , ":", max_ang, "\u212b" + , "\nThe min distance", gene, "structure df" , ":", min_ang, "\u212b") #-------------------------------- # call: combining_dfs_plotting() @@ -81,14 +86,22 @@ all_plot_dfs = combining_dfs_plotting(my_df_u , lig_dist_colname = LigDist_colname , lig_dist_cutoff = LigDist_cutoff) -merged_df2 = all_plot_dfs[[1]] -merged_df3 = all_plot_dfs[[2]] -merged_df2_comp = all_plot_dfs[[3]] -merged_df3_comp = all_plot_dfs[[4]] -merged_df2_lig = all_plot_dfs[[5]] -merged_df3_lig = all_plot_dfs[[6]] -merged_df2_comp_lig = all_plot_dfs[[7]] -merged_df3_comp_lig = all_plot_dfs[[8]] +merged_df2 = all_plot_dfs[[1]] +merged_df3 = all_plot_dfs[[2]] +#====================================================================== +# read other files +infilename_dynamut = paste0("~/git/Data/", drug, "/output/dynamut_results/", gene + , "_complex_dynamut_norm.csv") + +infilename_dynamut2 = paste0("~/git/Data/", drug, "/output/dynamut_results/dynamut2/", gene + , "_complex_dynamut2_norm.csv") + +infilename_mcsm_na = paste0("~/git/Data/", drug, "/output/mcsm_na_results/", gene + , "_complex_mcsm_na_norm.csv") + +dynamut_df = read.csv(infilename_dynamut) +dynamut2_df = read.csv(infilename_dynamut2) +mcsm_na_df = read.csv(infilename_mcsm_na) #################################################################### # Data for subcols barplot (~heatmpa) @@ -168,61 +181,6 @@ subcolsR_ps <- ColourPalleteMulti(subcols_df_ps, "duet_outcome", "my_grp_r") print(paste0("Colour palette generated for my_grp: ", length(subcols_ps), " colours")) print(paste0("Colour palette generated for my_grp_r: ", length(subcolsR_ps), " colours")) -#======================= -# Data for sub colours -# barplot: LIG -#======================= -cat("\nNo. of cols to select:", length(cols_to_select)) - -subcols_df_lig = merged_df3_lig[, cols_to_select] - -cat("\nNo of unique positions for LIG:" - , length(unique(subcols_df_lig$position))) - -# should be a factor -if (is.factor(subcols_df_lig$ligand_outcome)){ - cat("\nLigand_outcome is factor") - table(subcols_df_lig$ligand_outcome) -}else{ - cat("\nConverting ligand_outcome to factor") - subcols_df_lig$ligand_outcome = as.factor(subcols_df_lig$ligand_outcome) - table(subcols_df_lig$ligand_outcome) -} - -# should be -1 and 1 -min(subcols_df_lig$affinity_scaled) -max(subcols_df_lig$affinity_scaled) - -tapply(subcols_df_lig$affinity_scaled, subcols_df_lig$ligand_outcome, min) -tapply(subcols_df_lig$affinity_scaled, subcols_df_lig$ligand_outcome, max) - -# check unique values in normalised data -cat("\nNo. of unique values in affinity scaled, no rounding:" - , length(unique(subcols_df_lig$affinity_scaled))) - -# No rounding -my_grp_lig = subcols_df_lig$affinity_scaled; length(my_grp_lig) - -# Add rounding is to be used -n = 3 -subcols_df_lig$affinity_scaledR = round(subcols_df_lig$affinity_scaled, n) - -cat("\nNo. of unique values in duet scaled", n, "places rounding:" - , length(unique(subcols_df_lig$affinity_scaledR))) - -my_grp_lig_r = subcols_df_lig$affinity_scaledR # rounding - -# Add grp cols -subcols_df_lig$group_lig <- paste0(subcols_df_lig$ligand_outcome, "_", my_grp_lig, sep = "") -subcols_df_lig$group_ligR <- paste0(subcols_df_lig$ligand_outcome, "_", my_grp_lig_r, sep = "") - -# Call the function to create the palette based on the group defined above -subcols_lig <- ColourPalleteMulti(subcols_df_lig, "ligand_outcome", "my_grp_lig") -subcolsR_lig <- ColourPalleteMulti(subcols_df_lig, "ligand_outcome", "my_grp_lig_r") - -print(paste0("Colour palette generated for my_grp: ", length(subcols_lig), " colours")) -print(paste0("Colour palette generated for my_grp_r: ", length(subcolsR_lig), " colours")) - #################################################################### # Data for logoplots #################################################################### @@ -472,113 +430,6 @@ if (nrow(corr_ps_df3) == nrow(merged_df3) && nrow(merged_df3_comp) == check1) { , "\nGot: ", check1) } -#================================= -# Data for Correlation plots: LIG -#================================= -cat("\n==========================================" - , "\nCORR PLOTS data: LIG" - , "\n===========================================") - -df_lig = merged_df2_lig - -table(df_lig$ligand_outcome) - -#-------------------- -# adding log cols : NEW UNCOMMENT -#-------------------- -#df_lig$log10_or_mychisq = log10(df_lig$or_mychisq) -#df_lig$neglog_pval_fisher = -log10(df_lig$pval_fisher) - -##df_lig$log10_or_kin = log10(df_lig$or_kin) -##df_lig$neglog_pwald_kin = -log10(df_lig$pwald_kin) - -#---------------------------- -# columns for corr plots:PS -#---------------------------- -# subset data to generate pairwise correlations -cols_to_select = c("mutationinformation" - , "affinity_scaled" - #, "mutation_info_labels" - , "asa" - , "rsa" - , "rd_values" - , "kd_values" - , "log10_or_mychisq" - , "neglog_pval_fisher" - ##, "or_kin" - ##, "neglog_pwald_kin" - , "af" - ##, "af_kin" - , "ligand_outcome" - , drug) - -corr_data_lig = df_lig[, cols_to_select] - -dim(corr_data_lig) - -#-------------------------------------- -# assign nice colnames (for display) -#-------------------------------------- -my_corr_colnames = c("Mutation" - , "Ligand Affinity" - #, "Mutation class" - , "ASA" - , "RSA" - , "RD" - , "KD" - , "Log (OR)" - , "-Log (P)" - ##, "Adjusted (OR)" - ##, "-Log (P wald)" - , "MAF" - ##, "MAF_kin" - , "ligand_outcome" - , drug) - -length(my_corr_colnames) - -colnames(corr_data_lig) -colnames(corr_data_lig) <- my_corr_colnames -colnames(corr_data_lig) - -start = 1 -end = which(colnames(corr_data_lig) == drug); end # should be the last column -offset = 1 - -#============================= -# Corr data for plots: LIG -# big_df lig: ~ merged_df2_lig -#============================== -#corr_lig_df2 = corr_data_lig[start:(end-offset)] # without drug -corr_lig_df2 = corr_data_lig[start:end] -head(corr_lig_df2) - -#============================= -# Corr data for plots: LIG -# short_df lig: ~ merged_df3_lig -#============================== -corr_lig_df3 = corr_lig_df2[!duplicated(corr_lig_df2$Mutation),] - -na_or_lig = sum(is.na(corr_lig_df3$`Log (OR)`)) -check1_lig = nrow(corr_lig_df3) - na_or_lig - -if (nrow(corr_lig_df3) == nrow(merged_df3_lig) && nrow(merged_df3_comp_lig) == check1_lig) { - cat( "\nPASS: No. of rows for corr_lig_df3 match" - , "\nPASS: No. of OR values checked: " , check1_lig) -} else { - cat("\nFAIL: Numbers mismatch:" - , "\nExpected nrows: ", nrow(merged_df3_lig) - , "\nGot: ", nrow(corr_ps_df3_lig) - , "\nExpected OR values: ", nrow(merged_df3_comp_lig) - , "\nGot: ", check1_lig) -} - -# remove unnecessary columns -identical(corr_data_lig, corr_lig_df2) -identical(corr_data_ps, corr_ps_df2) - -#rm(df_ps, df_lig, corr_data_ps, corr_data_lig) - ######################################################################## # End of script ######################################################################## diff --git a/scripts/plotting/other_plots_combined.R b/scripts/plotting/other_plots_combined.R index d927808..3047f38 100644 --- a/scripts/plotting/other_plots_combined.R +++ b/scripts/plotting/other_plots_combined.R @@ -35,7 +35,7 @@ plot_dr_other_combined_labelled = paste0(plotdir,"/", dr_other_combined_labell #my_comparisons <- list( c(dr_muts_col, other_muts_col) ) my_comparisons <- list( c("DM", "OM") ) -my_ats = 22# axis text size +my_ats = 22 # axis text size my_als = 20 # axis label size my_fls = 20 # facet label size my_pts = 22 # plot title size @@ -45,12 +45,15 @@ my_pts = 22 # plot title size #=========== # Plot1: PS #=========== -my_stat_ps = compare_means(param_value~mutation_info, group.by = "param_type" - , data = df_lf_ps, paired = FALSE, p.adjust.method = "BH") +# my_stat_ps = compare_means(param_value~mutation_info +# , group.by = "param_type" +# , data = df_lf_ps +# , paired = FALSE +# , p.adjust.method = "BH") y_value = "param_value" -p1 = ggplot(df_lf_ps, aes(x = mutation_info +p1 = ggplot(lf_duet, aes(x = mutation_info , y = eval(parse(text=y_value)) )) + facet_wrap(~ param_type , nrow = 1 @@ -61,7 +64,7 @@ p1 = ggplot(df_lf_ps, aes(x = mutation_info geom_point(position = position_jitterdodge(dodge.width=0.01) , alpha = 0.5 , show.legend = FALSE - , aes(colour = factor(duet_outcome))) + + , aes(colour = duet_outcome)) + theme(axis.text.x = element_text(size = my_ats) , axis.text.y = element_text(size = my_ats , angle = 0 diff --git a/scripts/plotting/other_plots_data.R b/scripts/plotting/other_plots_data.R index df5c1e3..8fc9e00 100644 --- a/scripts/plotting/other_plots_data.R +++ b/scripts/plotting/other_plots_data.R @@ -5,21 +5,18 @@ ######################################################### #======================================================================= # working dir and loading libraries -getwd() -setwd("~/git/LSHTM_analysis/scripts/plotting") -getwd() +# getwd() +# setwd("~/git/LSHTM_analysis/scripts/plotting") +# getwd() -#source("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) +# 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" @@ -27,275 +24,515 @@ cols_to_select = c("mutation", "mutationinformation" merged_df3_short = merged_df3[, cols_to_select] -# write merged_df3 to generate structural figure -write.csv(merged_df3_short, "merged_df3_short.csv") +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") #======================================================================== -#%%%%%%%%%%%%%%%%%%% -# REASSIGNMENT: PS -#%%%%%%%%%%%%%%%%%%%% -df_ps = merged_df3 +# MOVE TO COMBINE or singular file for deepddg #============================ -# adding foldx scaled values +# adding deepddg scaled values # scale data b/w -1 and 1 #============================ -n = which(colnames(df_ps) == "ddg"); n +n = which(colnames(merged_df3) == "deepddg"); n -my_min = min(df_ps[,n]); my_min -my_max = max(df_ps[,n]); my_max +my_min = min(merged_df3[,n]); my_min +my_max = max(merged_df3[,n]); my_max -df_ps$foldx_scaled = ifelse(df_ps[,n] < 0 - , df_ps[,n]/abs(my_min) - , df_ps[,n]/my_max) +merged_df3$deepddg_scaled = ifelse(merged_df3[,n] < 0 + , merged_df3[,n]/abs(my_min) + , merged_df3[,n]/my_max) # sanity check -my_min = min(df_ps$foldx_scaled); my_min -my_max = max(df_ps$foldx_scaled); my_max +my_min = min(merged_df3$deepddg_scaled); my_min +my_max = max(merged_df3$deepddg_scaled); my_max if (my_min == -1 && my_max == 1){ - cat("PASS: foldx ddg successfully scaled b/w -1 and 1" - , "\nProceeding with assigning foldx outcome category") + cat("PASS: DeepDDG successfully scaled b/w -1 and 1" + #, "\nProceeding with assigning deep outcome category") + , "\n") }else{ - cat("FAIL: could not scale foldx ddg values" + cat("FAIL: could not scale DeepDDG ddg values" , "Aborting!") } -#================================ -# adding foldx outcome category -# ddg<0 = "Stabilising" (-ve) -#================================= +#======================================================================== +# cols to select -c1 = table(df_ps$ddg < 0) -df_ps$foldx_outcome = ifelse(df_ps$ddg < 0, "Stabilising", "Destabilising") -c2 = table(df_ps$ddg < 0) +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")] -if ( all(c1 == c2) ){ - cat("PASS: foldx outcome successfully created") -}else{ - cat("FAIL: foldx outcome could not be created. Aborting!") - exit() +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) #======================================================================= -# 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$ligand_outcome) +fact_cols = colnames(comb_df_s)[grepl( "_outcome|_info", colnames(comb_df_s) )] +fact_cols +lapply(comb_df_s[, fact_cols], class) +comb_df_s[,fact_cols] <- lapply(comb_df_s[,cols],as.factor) -# check -table(df_ps$mutation_info) +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$mutation_info) # further checks to make sure dr and other muts are indeed unique -dr_muts = df_ps[df_ps$mutation_info == dr_muts_col,] +dr_muts = comb_df_s[comb_df_s$mutation_info == dr_muts_col,] dr_muts_names = unique(dr_muts$mutation) -other_muts = df_ps[df_ps$mutation_info == other_muts_col,] +other_muts = comb_df_s[comb_df_s$mutation_info == other_muts_col,] other_muts_names = unique(other_muts$mutation) if ( table(dr_muts_names%in%other_muts_names)[[1]] == length(dr_muts_names) && table(other_muts_names%in%dr_muts_names)[[1]] == length(other_muts_names) ){ cat("PASS: dr and other muts are indeed unique") }else{ - cat("FAIL: dr adn others muts are NOT unique!") + cat("FAIL: dr and others muts are NOT unique!") quit() } +# pretty display names i.e. labels to reduce major code duplication later +foo_cnames = data.frame(colnames(comb_df_s)) +names(foo_cnames) <- "old_name" -#%%%%%%%%%%%%%%%%%%% -# REASSIGNMENT: LIG -#%%%%%%%%%%%%%%%%%%%% +stability_suffix <- paste0(delta_symbol, delta_symbol, "G") +flexibility_suffix <- paste0(delta_symbol, delta_symbol, "S") -df_lig = merged_df3_lig +lig_dn = paste0("Ligand distance (", angstroms_symbol, ")"); lig_dn +duet_dn = paste0("DUET ", stability_suffix); duet_dn +foldx_dn = paste0("FoldX ", stability_suffix); foldx_dn +deepddg_dn = paste0("Deepddg " , stability_suffix); deepddg_dn +mcsm_na_dn = paste0("mCSM-NA affinity ", stability_suffix); mcsm_na_dn +dynamut_dn = paste0("Dynamut ", stability_suffix); dynamut_dn +dynamut2_dn = paste0("Dynamut2 " , stability_suffix); dynamut2_dn +encom_ddg_dn = paste0("EnCOM " , stability_suffix); encom_ddg_dn +encom_dds_dn = paste0("EnCOM " , flexibility_suffix ); encom_dds_dn +sdm_dn = paste0("SDM " , stability_suffix); sdm_dn +mcsm_dn = paste0("mCSM " , stability_suffix ); mcsm_dn -# name tidying -df_lig$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" +# 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" - , "ligand_distance" - , "asa" - , "rsa" - , "rd_values" - , "kd_values") + , "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) + ) -df_wf_ps = df_ps[, cols_to_select_ps] +foo_cnames <- cbind(foo_cnames, colnames(comb_df_sl)) -pivot_cols_ps = cols_to_select_ps[1:5]; pivot_cols_ps +# some more pretty labels +table(comb_df_sl$mutation_info) -expected_rows_lf_ps = nrow(df_wf_ps) * (length(df_wf_ps) - length(pivot_cols_ps)) -expected_rows_lf_ps +levels(comb_df_sl$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 -df_lf_ps = gather(df_wf_ps, param_type, param_value, duet_scaled:kd_values, factor_key=TRUE) +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(df_lf_ps) == expected_rows_lf_ps){ - cat("PASS: long format data created for duet") +if (nrow(lf_duet) == expected_rows_lf){ + cat("\nPASS: long format data created for ", duet_dn) }else{ - cat("FAIL: long format data could not be created for duet") - exit() + cat("\nFAIL: long format data could not be created for duet") + quit() } -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$mutation_info) - ############################################################################ +#============== +# FoldX: LF +#============== +cols_to_select_foldx= c(static_cols_start, c("foldx_outcome", foldx_dn), static_cols_end) +wf_foldx = comb_df_sl[, cols_to_select_foldx] -#=========== -# LF data: LIG -#=========== -# keep similar dtypes cols together -cols_to_select_lig = c("mutationinformation", "mutation", "position", "mutation_info" - , "ligand_outcome" - - , "affinity_scaled" - #, "ligand_distance" - , "asa" - , "rsa" - , "rd_values" - , "kd_values") +pivot_cols_foldx = cols_to_select_foldx[1: (length(static_cols_start) + 1)]; pivot_cols_foldx -df_wf_lig = df_lig[, cols_to_select_lig] +expected_rows_lf = nrow(wf_foldx) * (length(wf_foldx) - length(pivot_cols_foldx)) +expected_rows_lf -pivot_cols_lig = cols_to_select_lig[1:5]; pivot_cols_lig +# LF data: duet +print("TESTXXXXXXXXXXXXXXXXXXXXX---------------------->>>>") +lf_foldx <<- gather(wf_foldx + , key = param_type + , value = param_value + , all_of(foldx_dn):tail(static_cols_end,1) + , factor_key = TRUE) -expected_rows_lf_lig = nrow(df_wf_lig) * (length(df_wf_lig) - length(pivot_cols_lig)) -expected_rows_lf_lig - -# LF data: foldx -df_lf_lig = gather(df_wf_lig, param_type, param_value, affinity_scaled:kd_values, factor_key=TRUE) - -if (nrow(df_lf_lig) == expected_rows_lf_lig){ - cat("PASS: long format data created for foldx") +if (nrow(lf_foldx) == expected_rows_lf){ + cat("\nPASS: long format data created for ", foldx_dn) }else{ - cat("FAIL: long format data could not be created for foldx") - exit() + cat("\nFAIL: long format data could not be created for duet") + quit() } -# assign pretty labels: param_type -levels(df_lf_lig$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) - ############################################################################ +#============== +# 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] -# 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) +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() +} +############################################################################ +# # clear excess variables +# rm(all_plot_dfs +# , cols_dynamut2_df +# , cols_mcsm_df +# , cols_mcsm_na_df +# , comb_df +# , corr_data_ps +# , corr_ps_df3 +# , df_lf_ps +# , foo +# , foo_cnames +# , gene_metadata +# , logo_data +# , logo_data_or_mult +# , logo_data_plot +# , logo_data_plot_logor +# , logo_data_plot_or +# , my_data_snp +# , my_df +# , my_df_u +# , ols_mcsm_df +# , other_muts +# , pd_df +# , subcols_df_ps +# , tab_mt +# , wide_df_logor +# , wide_df_logor_m +# , wide_df_or +# , wide_df_or_mult +# , wt) +# +# +# rm(c3, c4, check1 +# , cols_check +# , cols_to_select +# , cols_to_select_deepddg +# , cols_to_select_duet +# , cols_to_select_dynamut +# , cols_to_select_dynamut2 +# , cols_to_select_encomddg +# , cols_to_select_encomdds +# , cols_to_select_mcsm +# , cols_to_select_mcsm_na +# , cols_to_select_sdm)