#!/usr/bin/env Rscript       
#source("~/git/LSHTM_analysis/config/katg.R")
#source("~/git/LSHTM_analysis/scripts/plotting/get_plotting_dfs.R")

#=======
# output
#=======
outdir_images = paste0("~/git/Writing/thesis/images/results/", tolower(gene), "/")
outdir_stats = paste0(outdir_images,"stats/") 
cat("\nOutput dir for stats:", outdir_stats)
###################################################################
geneL_normal  = c("pnca")
#geneL_na      = c("gid", "rpob")
geneL_na_v2   = c("gid")
geneL_ppi2    = c("alr", "embb", "katg", "rpob")
geneL_both    = c("rpob")

if (tolower(gene)%in%geneL_na_v2) {
  gene_colnames = c("mcsm_na_affinity", "mcsm_na_outcome")
}

if (tolower(gene)%in%geneL_ppi2) {
  gene_colnames = c("mcsm_ppi2_affinity", "mcsm_ppi2_outcome")
}

#from plotting_globals()
LigDist_colname
ppi2Dist_colname 
naDist_colname

delta_symbol #delta_symbol = "\u0394"; delta_symbol
angstroms_symbol

cat("\nAffinity Distance colnames:", length(affinity_dist_colnames)
    , "\nThese are:", affinity_dist_colnames)
#===========
# Data used
#===========
df3 = merged_df3

cols_to_output = c("position"
                   , "sensitivity"
                   , "mutationinformation"
                   , affinity_dist_colnames[1]
                   , "ligand_affinity_change"
                   , "ligand_outcome"
                   , "mmcsm_lig"
                   , "mmcsm_lig_outcome"
                   , affinity_dist_colnames[2]
                   # #, affinity_dist_colnames[3]
                   # , "mcsm_na_affinity"
                   # , "mcsm_na_outcome"
                   # #, "mcsm_nca_affinity"
                   # #, "mcsm_nca_outcome"
                   , gene_colnames
                   , "maf"
                   , "or_mychisq"
                   , "pval_fisher")

cols_to_output
df3_output = df3[, cols_to_output]
colnames(df3_output)
cat("\nSelecting columns:", length(colnames(df3_output)))
#===============================================
# Add COLS and rounding: adjusted P-values + MAF
#==============================================
#-----------------------------
# adjusted P-values
#-----------------------------
# add cols: p_adj_fdr and signif_fdr
df3_output$p_adj_fdr = p.adjust(df3_output$pval_fisher, method = "fdr")
df3_output$signif_fdr = df3_output$p_adj_fdr
df3_output = dplyr::mutate(df3_output
                           , signif_fdr = case_when(signif_fdr == 0.05 ~ "."
                                                    , signif_fdr <=0.0001 ~ '****'
                                                    , signif_fdr <=0.001 ~ '***'
                                                    , signif_fdr <=0.01 ~ '**'
                                                    , signif_fdr <0.05 ~ '*'
                                                    , TRUE ~ 'ns'))
# rounding
df3_output$or_mychisq = round(df3_output$or_mychisq,2)
df3_output$p_adj_fdr  = round(df3_output$p_adj_fdr,2)
head(df3_output)

#----------
# MAF (%)
#----------
# add col maf_percent
df3_output$maf_percent = df3_output$maf*100

# rounding
df3_output$maf_percent = round(df3_output$maf_percent,2)
head(df3_output$af); head(df3_output$maf);head(df3_output$maf_percent)

#----------
# P-value
#----------
df3_output$pval_fisher = round(df3_output$pval_fisher,2)

class(df3_output)
head(df3_output)

####################################
# Appendix: ligand affinity
####################################
df_lig = df3_output[df3_output[[LigDist_colname]]<DistCutOff,]

cols_to_output_lig = c("position"
                       , "sensitivity"
                       , "mutationinformation"
                       , LigDist_colname
                       , "ligand_affinity_change"
                       , "ligand_outcome"
                       , "mmcsm_lig"
                       , "mmcsm_lig_outcome"
                       , "maf_percent"
                       , "or_mychisq"
                       , "pval_fisher"
                       , "p_adj_fdr"
                       , "signif_fdr")
# select cols
Out_df_lig = df_lig[, cols_to_output_lig]

# sort df by OR and then MAF: highest OR and highest MAF
#Out_df_ligS1 = Out_df_lig[order(Out_df_lig$or_mychisq, decreasing = T), ]
Out_df_ligS = Out_df_lig[order(-Out_df_lig$or_mychisq, Out_df_lig$maf_percent), ]

#head(Out_df_ligS1); tail(Out_df_ligS1)
head(Out_df_ligS); tail(Out_df_ligS)

colsNames_to_output_lig = c("position"
                            , "sensitivity"
                            , "Mutation"
                            , paste0("Lig-Dist (", angstroms_symbol, ")")
                            , "mCSM-ligand affinity"
                            , "mCSM ligand_outcome"
                            , "mmCSM-ligand affinity"
                            , "mmCSM ligand_outcome"
                            , paste0("MAF ","(%)")
                            , "Odds Ratio"
                            , "P-value"
                            , "Adjusted P-value"
                            , "Adjusted P-value significance")

colnames(Out_df_ligS) = colsNames_to_output_lig
head(Out_df_ligS)

# ADD: active site annot
nrow(Out_df_ligS)
Out_df_ligS$drug_site = ifelse(Out_df_ligS$position%in%aa_pos_drug, "drug", "no")
table(Out_df_ligS$drug_site)

Out_df_ligS$heme_site = ifelse(Out_df_ligS$position%in%aa_pos_hem, "heme", "no")
table(Out_df_ligS$heme_site)

#--------------------
# write output file: KS test within grpup
#----------------------
Out_ligT = paste0(outdir_stats
                  , tolower(gene)
                  , "_lig_muts.csv")

cat("Output of Ligand muts:", Out_ligT )
write.csv(Out_df_ligS, Out_ligT, row.names = FALSE)

########################################################################
####################################
# Appendix: NA/PPI2 affinity
# naDist_colname
# ppi2Dist_colname
####################################
# Filtered data
#df_nca = df3_output[df3_output[[naDist_colname]]<DistCutOff,]
df_nca = df3_output[df3_output[[ppi2Dist_colname]]<DistCutOff,]

# select cols
cols_to_output_nca = c("position"
                        , "sensitivity"
                        , "mutationinformation"
                        #, naDist_colname
                        , ppi2Dist_colname
                        , gene_colnames
                        , "maf_percent"
                        , "or_mychisq"
                        , "pval_fisher"
                        , "p_adj_fdr"
                        , "signif_fdr")

# extract output cols
Out_df_nca = df_nca[, cols_to_output_nca]

# sort df by OR and then MAF: Highest OR and Highest MAF
#Out_df_ncaS = Out_df_nca[order(Out_df_nca$or_mychisq, decreasing = T), ]
Out_df_ncaS = Out_df_nca[order(-Out_df_nca$or_mychisq, Out_df_nca$maf_percent), ]

colsNames_to_output_nca = c("position"
                             , "sensitivity"
                             , "Mutation"
                            
                             # , paste0("NA-Dist (", angstroms_symbol, ")")
                             # , paste0("mCSM-NA (", delta_symbol,delta_symbol,"G)")
                             # , "mCSM-NA outcome"
                            
                            , paste0("PPI-Dist (", angstroms_symbol, ")")
                            , paste0("mCSM-PPI (", delta_symbol,delta_symbol,"G)")
                            , "mCSM-PPI outcome"
                            
                             , paste0("MAF ","(%)")
                             , "Odds Ratio"
                             , "P-value"
                             , "Adjusted P-value"
                             , "Adjusted P-value significance")

colnames(Out_df_ncaS) = colsNames_to_output_nca
Out_df_ncaS

# ADD: active site annot
nrow(Out_df_ncaS)
Out_df_ncaS$drug_site = ifelse(Out_df_ncaS$position%in%aa_pos_drug, "drug", "no")
table(Out_df_ncaS$drug_site)

Out_df_ncaS$heme_site = ifelse(Out_df_ncaS$position%in%aa_pos_hem, "heme", "no")
table(Out_df_ncaS$heme_site)

#--------------------
# write output file: KS test within grpup
#----------------------
Out_ncaT = paste0(outdir_stats
                   , tolower(gene)
                   , "_ppi2_muts.csv")

cat("Output of NA muts:", Out_ncaT )
write.csv(Out_df_ncaS, Out_ncaT, row.names = FALSE)

#################################################################################
#################################################################################
#################################################################################
##########################################################
# higest or/maf and stability effects
###########################################################
# convert to percet
df3$maf_percent = df3$maf*100

cols_to_output_effects = c("position"
                   , "sensitivity"
                   , "mutationinformation"
                   , "avg_stability"
                   , "avg_stability_outcome"
                   , affinity_dist_colnames[1]
                   , "avg_lig_affinity"
                   , "avg_lig_affinity_outcome"
                   , affinity_dist_colnames[2]
                   , gene_colnames
                   , "maf_percent"
                   , "or_mychisq"
                   , "pval_fisher")

df3_effects = df3[, cols_to_output_effects]
nrow(df3_effects); ncol(df3_effects)

# add cols: p_adj_fdr and signif_fdr
df3_effects$p_adj_fdr = p.adjust(df3_effects$pval_fisher, method = "fdr")
df3_effects$signif_fdr = df3_effects$p_adj_fdr
df3_effects = dplyr::mutate(df3_effects
                           , signif_fdr = case_when(signif_fdr == 0.05 ~ "."
                                                    , signif_fdr <=0.0001 ~ '****'
                                                    , signif_fdr <=0.001 ~ '***'
                                                    , signif_fdr <=0.01 ~ '**'
                                                    , signif_fdr <0.05 ~ '*'
                                                    , TRUE ~ 'ns'))
# rounding
df3_effects$or_mychisq = round(df3_effects$or_mychisq,2)
df3_effects$p_adj_fdr  = round(df3_effects$p_adj_fdr,2)
head(df3_effects)

#-------------------
# Highest OR and MAF
#-------------------
mut_hor  = df3_effects[df3_effects$or_mychisq%in%(max(df3_effects$or_mychisq, na.rm = T)),]
mut_hmaf = df3_effects[df3_effects$maf_percent%in%(max(df3_effects$maf_percent, na.rm = T)),]

if (identical(colnames(mut_hor), colnames(mut_hmaf)) ){
  
  # add cols
  mut_hor$mutational_effect  = "Mutation with highest OR"
  mut_hmaf$mutational_effect = "Most frequent mutation"
  cat("\nPass: or and maf")
}else{
  quit("Abort: colnames or and maf mismatch")
}

#-------------------
# Avg stability
# most DD/SS: average stability
#-------------------
mut_h_avs_dd = df3_effects[df3_effects$avg_stability%in%(min(df3_effects$avg_stability, na.rm = T)),]
mut_h_avs_ss = df3_effects[df3_effects$avg_stability%in%(max(df3_effects$avg_stability, na.rm = T)),]

if (identical(colnames(mut_h_avs_dd), colnames(mut_h_avs_ss)) ){
  
  # add cols
  mut_h_avs_dd$mutational_effect = "Most Destabilising for protomer"
  mut_h_avs_ss$mutational_effect = "Most Stabilising for protomer"
  
  cat("\nPass : avg stability")
}else{
  quit("Abort: colnames stability mismatch")
}

#-------------------
# Filtered columns
# most DD/SS: ligand
# FIXME DUBIOUS as min and max can be both negative
#-------------------
df3_effects_lig = df3_effects[df3_effects[[LigDist_colname]]<DistCutOff,]
nrow(df3_effects_lig)

mut_h_lig_dd = df3_effects_lig[df3_effects_lig$avg_lig_affinity%in%(min(df3_effects_lig$avg_lig_affinity, na.rm = T)),]
mut_h_lig_ss = df3_effects_lig[df3_effects_lig$avg_lig_affinity%in%(max(df3_effects_lig$avg_lig_affinity, na.rm = T)),]

if (identical(colnames(mut_h_lig_dd), colnames(mut_h_lig_ss)) ){
  
  # add cols
  mut_h_lig_dd$mutational_effect = "Most Destabilising for Ligand affinity"
  mut_h_lig_ss$mutational_effect = "CAUTION: Most DE/Stabilising for Ligand affinity"
  
  cat("\nPass: avg ligand affinity")
}else{
  quit("Abort: colnames lig mismatch")
}

#-------------------
# Filtered columns
# most DD/SS: NA
#-------------------
if (tolower(gene)%in%geneL_na_v2 ){
  
  df3_effects_na = df3_effects[df3_effects[[naDist_colname]]<DistCutOff,]
  nrow(df3_effects_na)
  mut_h_na_dd = df3_effects_na[df3_effects_na$mcsm_na_affinity%in%(min(df3_effects_na$mcsm_na_affinity, na.rm = T)),]
  mut_h_na_ss = df3_effects_na[df3_effects_na$mcsm_na_affinity%in%(max(df3_effects_na$mcsm_na_affinity, na.rm = T)),]
  
  # add cols
  mut_h_na_dd$mutational_effect = "Most Destabilising for NA affinity"
  mut_h_na_ss$mutational_effect = "Most Stabilising for NA affinity"
  
  if (identical(colnames(mut_h_na_dd), colnames(mut_h_na_ss)) ){
    cat("\nPass 1: NCA")
  }else{
    quit("Abort: colnames nca mismatch")
  }
  
  if (identical(colnames(mut_h_na_dd), colnames(mut_h_lig_dd)) ){
    cat("\nPass 2: NCA")
  }else{
    quit("Abort: colnames ppi2 mismatch")
  }
  #combine
  gene_aff_combined = rbind(mut_h_na_dd, mut_h_na_ss)
  
}

#-------------------
# Filtered columns
# most DD/SS: ppi2
#-------------------
if (tolower(gene)%in%geneL_ppi2 ){

  df3_effects_ppi2 = df3_effects[df3_effects[[ppi2Dist_colname]]<DistCutOff,]
  nrow(df3_effects_ppi2)

  mut_h_ppi2_dd = df3_effects_ppi2[df3_effects_ppi2$mcsm_ppi2_affinity%in%(min(df3_effects_ppi2$mcsm_ppi2_affinity, na.rm = T)),]
  mut_h_ppi2_ss = df3_effects_ppi2[df3_effects_ppi2$mcsm_ppi2_affinity%in%(max(df3_effects_ppi2$mcsm_ppi2_affinity, na.rm = T)),]
  
  # add cols
  mut_h_ppi2_dd$mutational_effect = "Most Destabilising for PPI affinity"
  mut_h_ppi2_ss$mutational_effect = "Most Stabilising for PPI affinity"
  
  if (identical(colnames(mut_h_ppi2_dd), colnames(mut_h_ppi2_ss)) ){
    cat("\nPass 1: ppi2")
  }else{
    quit("Abort: colnames ppi2 mismatch")
  }
  
  if (identical(colnames(mut_h_ppi2_dd), colnames(mut_h_lig_dd)) ){
    cat("\nPass 2 : ppi2")
  }else{
    quit("Abort: colnames ppi2 mismatch")
  }
  #combine
  gene_aff_combined = rbind(mut_h_ppi2_dd, mut_h_ppi2_ss)
  
}

#============
# final combine
#============
if ( identical(colnames(mut_hor), colnames(mut_h_lig_dd)) ){
  cat("PASS: all")
  combined_table = rbind(mut_hor, mut_hmaf,
                         mut_h_avs_dd, mut_h_avs_ss,
                         mut_h_lig_dd, mut_h_lig_ss,
                         gene_aff_combined)
  cat("\nCombined table dim:", "\nnrow:", nrow(combined_table), "\nncol:", ncol(combined_table))
}else{
  quit("Abort: colnames ppi2 mismatch")
}

# Assign pretty colnames
colnames(combined_table)
colsNames_combined_table = c("position"
                            , "sensitivity"
                            , "Mutation"
                            , paste0("Avg stability (", delta_symbol,delta_symbol,"G)")
                            , "avg stability outcome"
                            
                            , paste0("Lig-Dist (", angstroms_symbol, ")")
                            , "Avg ligand affinity"
                            , "Ligand affinity outcome"
                            
                            # , paste0("NA-Dist (", angstroms_symbol, ")")
                            # , paste0("mCSM-NA (", delta_symbol,delta_symbol,"G)")
                            # , "mCSM-NA outcome"
                            
                            , paste0("PPI-Dist (", angstroms_symbol, ")")
                            , paste0("mCSM-PPI (", delta_symbol,delta_symbol,"G)")
                            , "mCSM-PPI outcome"
                            
                            , paste0("MAF ","(%)")
                            , "Odds Ratio"
                            , "P-value"
                            , "Adjusted P-value"
                            , "Adjusted P-value significance"
                            , "Mutational effect")

if ( length(colnames(combined_table)) == length(colsNames_combined_table) ) {
  cat("Assiging pretty colnames for output")
  colnames(combined_table) <- colsNames_combined_table
  #colnames(combined_table)
}else{
  stop("\nAbort: No. of cols mismatch. Cannot assign pretty colnames for output")
}


nrow(combined_table)
combined_table$drug_site = ifelse(combined_table$position%in%aa_pos_drug, "drug", "no")
table(combined_table$drug_site)

combined_table$heme_site = ifelse(combined_table$position%in%aa_pos_hem, "heme", "no")
table(combined_table$heme_site)

#--------------------
# write output file: KS test within grpup
#----------------------
Out_combined_effectsT = paste0(outdir_stats
                  , tolower(gene)
                  , "_mut_effects.csv")

cat("Output of effects:", Out_combined_effectsT )
write.csv(combined_table, Out_combined_effectsT, row.names = FALSE)