generated subcolour bps for PS

scripts/plotting/barplots_subcolours_PS.R

@@ -3,105 +3,71 @@ setwd("~/git/LSHTM_analysis/scripts/plotting")
 getwd()
 
 #########################################################
-# TASK:
+# TASK: output barplot by position with each bar coloured by 
-
+# its stability value and NO coloured positions indicated
 #########################################################
+#=======================================================================
 
-########################################################################
+############################################################
-# 		Installing and loading required packages and functions		   #	
+# 1: Installing and loading required packages and functions
-########################################################################
+#############################################################
 
-source("Header_TT.R")
+#source("Header_TT.R")
+library(ggplot2)
+library(data.table)
 source("barplot_colour_function.R")
+source("plotting_data.R")
 
-########################################################################
+# should return the following dfs, directories and variables
-#		 Read file: call script for combining df for PS			   	   #
+# mut_pos_cols
-########################################################################
+# my_df
-#?????????????
+# my_df_u
-#
+# my_df_u_lig
+# dup_muts
+
+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(my_df, upos, dup_muts, my_df_u_lig)
 ########################################################
-#%% variable assignment: input and output paths & filenames
-drug = "pyrazinamide"
-gene = "pncA"
-gene_match = paste0(gene,"_p.")
-cat(gene_match)
-
-#=============
-# directories
-#=============
-datadir = paste0("~/git/Data")
-indir = paste0(datadir, "/", drug, "/input")
-outdir = paste0("~/git/Data", "/", drug, "/output")
-
-#======
-# input
-#======
-#in_filename = "mcsm_complex1_normalised.csv"
-in_filename_params = paste0(tolower(gene), "_all_params.csv") 
-infile_params = paste0(outdir, "/", in_filename_params)
-cat(paste0("Input file:", infile_params) )
 
 #=======
 # output
 #=======
-subcols_bp_duet = "barplot_subcols_DUET.svg"
+print(paste0("plot will be in:", plotdir))
-outPlot_subcols_bp_duet  =  paste0(outdir, "/plots/", subcols_bp_duet)
+bp_subcols_duet = "barplot_coloured_PS.svg"
+plot_bp_subcols_duet = paste0(plotdir, "/", bp_subcols_duet)
 
-#%%===============================================================
-###########################
-# Read file: struct params
-###########################
-cat("Reading struct params including mcsm:", in_filename_params)
-
-my_df = read.csv(infile_params
-                 #, stringsAsFactors = F
-                 , header = T)
-
-cat("Input dimensions:", dim(my_df)) 
-
-# clear variables
-rm(in_filename_params, infile_params)
-
-# quick checks
-colnames(my_df)
-str(my_df)
-
-# check for duplicate mutations
-if ( length(unique(my_df$mutationinformation)) != length(my_df$mutationinformation)){
-  cat(paste0("CAUTION:", " Duplicate mutations identified"
-             , "\nExtracting these..."))
-  dup_muts = my_df[duplicated(my_df$mutationinformation),]
-  dup_muts_nu = length(unique(dup_muts$mutationinformation))
-  cat(paste0("\nDim of duplicate mutation df:", nrow(dup_muts)
-             , "\nNo. of unique duplicate mutations:", dup_muts_nu
-             , "\n\nExtracting df with unique mutations only"))
-  my_df_u = my_df[!duplicated(my_df$mutationinformation),]
-}else{
-  cat(paste0("No duplicate mutations detected"))
-  my_df_u = my_df
-}
-
-upos = unique(my_df_u$position)
-cat("Dim of clean df:"); cat(dim(my_df_u))
-cat("\nNo. of unique mutational positions:"); cat(length(upos))
-
-########################################################################
-#               end of data extraction and cleaning for plots          #
-########################################################################
 #===================
 # Data for plots
 #===================
 # REASSIGNMENT as necessary
 df  = my_df_u
 
-rm(my_df)
-
 # sanity checks
+str(df)
 upos = unique(df$position)
 
 # should be a factor
-is.factor(my_df$duet_outcome)
+if (is.factor(df$duet_outcome)){
-#[1] TRUE
+  print("duet_outcome is factor")
+}else{
+  print("convert duet_outcome to factor")
+  df$duet_outcome = as.factor(df$duet_outcome)
+}
+
+is.factor(df$duet_outcome)
 
 table(df$duet_outcome)
 
@@ -167,12 +133,7 @@ my_yaxts = 15
 
 #******************
 # generate plot: NO axis colours
-# no ordering of x-axis
 #******************
-# plot name and location
-print(paste0("plot will be in:", outdir))
-bp_subcols_duet = "barplot_coloured_PS.svg"
-plot_bp_subcols_duet = paste0(outdir, "/plots/", bp_subcols_duet) 
 print(paste0("plot name:", plot_bp_subcols_duet))
 
 svg(plot_bp_subcols_duet, width = 26, height = 4)
@@ -192,7 +153,8 @@ outPlot = g +
                                       , vjust = 0)
          , axis.title.x = element_text(size = my_xaxts)
          , axis.title.y = element_text(size = my_yaxts ) ) +
-  labs(title = my_title
+  labs(title = ""
+       #title = my_title
        , x = "position"
        , y = "Frequency")
        

scripts/plotting/barplots_subcolours_aa_PS.R

@@ -75,6 +75,15 @@ my_axis_colours = mut_pos_cols$lab_fg
 
 # now clear mut_pos_cols
 rm(mut_pos_cols)
+
+#=======
+# output
+#=======
+# plot name and location
+print(paste0("plot will be in:", plotdir))
+bp_aa_subcols_duet = "barplot_acoloured_PS.svg"
+plot_bp_aa_subcols_duet = paste0(plotdir, "/", bp_aa_subcols_duet)
+
 #=======================================================================
 #================
 # Data for plots
@@ -125,7 +134,13 @@ snp_count = sort(unique(snpsBYpos_df$snpsBYpos))
 
 # sanity checks
 # should be a factor
-df$duet_outcome = as.factor(df$duet_outcome)
+if (is.factor(df$duet_outcome)){
+  print("duet_outcome is factor")
+}else{
+  print("convert duet_outcome to factor")
+  df$duet_outcome = as.factor(df$duet_outcome)
+}
+
 is.factor(df$duet_outcome)
 
 table(df$duet_outcome)
@@ -189,14 +204,6 @@ my_yats = 18
 #******************
 # generate plot: with axis colours
 #******************
-# plot name and location
-# outdir/ (should be imported from reading file)
-plotdir = paste0(outdir,  "/",  "plots") #should be imported from reading file
-print(paste0("plot will be in:", plotdir))
-
-bp_aa_subcols_duet = "barplot_acoloured_PS.svg"
-plot_bp_aa_subcols_duet = paste0(outdir, "/plots/", bp_aa_subcols_duet)
-
 print(paste0("plot name:", plot_bp_aa_subcols_duet))
 
 svg(plot_bp_aa_subcols_duet, width = 26, height = 4)
@@ -230,6 +237,7 @@ outPlot = g +
          , axis.title.y = element_text(size = my_yals ) 
          , axis.ticks.x = element_blank()) +
   labs(title = ""
+       #title = my_title
        , x = "position"
        , y = "Frequency")
 

scripts/plotting/running_plot_scripts

@@ -3,20 +3,75 @@
 #========
 
 #=======================
-#1) plotting_data.R
+plotting_data.R
 #=======================
 ??? update how to run
 
 #=======================
-# mcsm_mean_stability.R
+mcsm_mean_stability.R
 #=======================
 ??? update how to run
 # input: calls the "plotting_data.R"
 # output: <gene>_mean_stability.csv
 
 #======================
-# replaceBfactor_pdb.R
+replaceBfactor_pdb.R
 #=======================
 # input: 2 files; pdb file and <gene>_mean_stability.csv (output from "mcsm_mean_stability.R")
 # output: 2 pdb files with bfactors replaced with 1) mean_duet and 2) mean_affinity values
 
+#=======================
+basic_barplots_PS.R
+#=======================
+# input: calls the "plotting_data.R"
+output plots: 2 svgs
+    1) basic_barplot_PS.svg
+    2) position_count_PS.svg
+    
+#=======================    
+barplot_colour_function.R"
+#=======================  
+function that generates stability coloures (red to blue) based on the
+corresponding value. It is sourced by other scripts.
+
+#=======================
+subcols_axis_PS.R
+# assigns colours to position numebers with bg and fg colours
+# sourced by other scripts.
+#=======================
+# input: calls the "plotting_data.R"
+# sourced by other scripts, no output perse!
+   
+#=======================
+barplots_subcolours_aa_PS.R
+# barplot coloured by stability and position numbers coloured by active site colours
+#=======================
+# input: calls the "subcols_axis.R" and "barplot_colour_function.R"
+# output plots: 1 svg
+    1) barplot_acoloured_PS.svg
+
+#=======================
+barplots_subcolours_PS.R
+# same as the above script expcet no positional colouring
+#=======================
+# input: calls the "plotting_data.R" and "barplot_colour_function.R"
+# output plots: 1 svg
+    1) barplot_coloured_PS.svg
+
+
+    
+########################################################################
+#                              ligand plots
+########################################################################
+
+#=======================
+basic_barplots_LIG.R
+# generates two basic barplots 
+# No. of stabilsing and destabilising mutations for ligand_distance<10A
+# No. of SNPs per site
+#=======================
+# input: calls the "plotting_data.R"
+# output plots: 2 svgs
+    1) basic_barplot_LIG.svg
+    2) position_count_LIg.svg
+