tanu/LSHTM_analysis
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

renamed file to denote corr adjusted and plain

Browse source
This commit is contained in:
Tanushree Tunstall 2020-09-17 16:35:35 +01:00
parent fb0646373b
commit 1b5280145b
2 changed files with 95 additions and 95 deletions
Download patch file Download diff file Expand all files Collapse all files

87
scripts/plotting/corr_plots.R
View file

@ -14,6 +14,7 @@ getwd()
source("Header_TT.R") source("Header_TT.R")
require(cowplot) require(cowplot)
source("combining_dfs_plotting.R") source("combining_dfs_plotting.R")
source("my_pairs_panel.R") # with lower panel turned off
# should return the following dfs, directories and variables # should return the following dfs, directories and variables
# PS combined: # PS combined:
@ -77,33 +78,76 @@ rm( merged_df2, merged_df2_comp, merged_df2_lig, merged_df2_comp_lig, my_df_u, m
# end of data extraction and cleaning for plots # # end of data extraction and cleaning for plots #
######################################################################## ########################################################################
#=========================== #============================
# Data for Correlation plots:PS # adding foldx scaled values
#=========================== # scale data b/w -1 and 1
table(df_ps$duet_outcome) #============================
n = which(colnames(df_ps) == "ddg"); n
my_min = min(df_ps[,n]); my_min
my_max = max(df_ps[,n]); my_max
df_ps$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()
}
table(df_ps$foldx_outcome)
#======================
# adding log cols # adding log cols
#======================
df_ps$log10_or_mychisq = log10(df_ps$or_mychisq) df_ps$log10_or_mychisq = log10(df_ps$or_mychisq)
df_ps$neglog_pval_fisher = -log10(df_ps$pval_fisher) df_ps$neglog_pval_fisher = -log10(df_ps$pval_fisher)
df_ps$log10_or_kin = log10(df_ps$or_kin) df_ps$log10_or_kin = log10(df_ps$or_kin)
df_ps$neglog_pwald_kin = -log10(df_ps$pwald_kin) df_ps$neglog_pwald_kin = -log10(df_ps$pwald_kin)
#===========================
# Data for Correlation plots:PS
#===========================
# subset data to generate pairwise correlations # subset data to generate pairwise correlations
cols_to_select = c("duet_scaled" cols_to_select = c("duet_scaled"
, "foldx_scaled"
, "log10_or_mychisq" , "log10_or_mychisq"
, "neglog_pval_fisher" , "neglog_pval_fisher"
, "log10_or_kin"
, "neglog_pwald_kin"
, "af" , "af"
, "duet_outcome" , "duet_outcome"
, drug) , drug)
corr_data_ps = df_ps[, cols_to_select] corr_data_ps = df_ps[cols_to_select]
dim(corr_data_ps) dim(corr_data_ps)
@ -113,12 +157,11 @@ dim(corr_data_ps)
# assign nice colnames (for display) # assign nice colnames (for display)
my_corr_colnames = c("DUET" my_corr_colnames = c("DUET"
, "Foldx"
, "Log(OR)" , "Log(OR)"
, "-Log(P)" , "-Log(P)"
, "Log(OR adjusted)"
, "-Log(P wald)"
, "AF" , "AF"
, "duet_outcome" , "duet_outcome"
@ -147,7 +190,8 @@ head(my_corr_ps)
cat("Corr plot PS:", plot_corr_ps) cat("Corr plot PS:", plot_corr_ps)
svg(plot_corr_ps, width = 15, height = 15) svg(plot_corr_ps, width = 15, height = 15)
OutPlot1 = pairs.panels(my_corr_ps[1:(length(my_corr_ps)-1)] #OutPlot1 = pairs.panels([1:(length(my_corr_ps)-1)]
OutPlot1 = my_pp(my_corr_ps[1:(length(my_corr_ps)-1)]
, method = "spearman" # correlation method , method = "spearman" # correlation method
, hist.col = "grey" ##00AFBB , hist.col = "grey" ##00AFBB
, density = TRUE # show density plots , density = TRUE # show density plots
@ -162,10 +206,11 @@ OutPlot1 = pairs.panels(my_corr_ps[1:(length(my_corr_ps)-1)]
#, alpha = .05 #, alpha = .05
#, points(pch = 19, col = c("#f8766d", "#00bfc4")) #, points(pch = 19, col = c("#f8766d", "#00bfc4"))
, cex = 3 , cex = 3
, cex.axis = 2.5 , cex.axis = 1.5
, cex.labels = 2.1 , cex.labels = 2.1
, cex.cor = 1 , cex.cor = 1
, smooth = F , smooth = F
) )
print(OutPlot1) print(OutPlot1)
@ -190,9 +235,6 @@ cols_to_select = c("affinity_scaled"
, "log10_or_mychisq" , "log10_or_mychisq"
, "neglog_pval_fisher" , "neglog_pval_fisher"
, "log10_or_kin"
, "neglog_pwald_kin"
, "af" , "af"
, "ligand_outcome" , "ligand_outcome"
@ -209,9 +251,6 @@ my_corr_colnames = c("Ligand Affinity"
, "Log(OR)" , "Log(OR)"
, "-Log(P)" , "-Log(P)"
, "Log(OR adjusted)"
, "-Log(P wald)"
, "AF" , "AF"
, "ligand_outcome" , "ligand_outcome"
@ -237,7 +276,8 @@ head(my_corr_lig)
cat("Corr LIG plot:", plot_corr_lig) cat("Corr LIG plot:", plot_corr_lig)
svg(plot_corr_lig, width = 15, height = 15) svg(plot_corr_lig, width = 15, height = 15)
OutPlot2 = pairs.panels(my_corr_lig[1:(length(my_corr_lig)-1)] #OutPlot2 = pairs.panels(my_corr_lig[1:(length(my_corr_lig)-1)]
OutPlot2 = my_pp(my_corr_lig[1:(length(my_corr_lig)-1)]
, method = "spearman" # correlation method , method = "spearman" # correlation method
, hist.col = "grey" ##00AFBB , hist.col = "grey" ##00AFBB
, density = TRUE # show density plots , density = TRUE # show density plots
@ -252,7 +292,7 @@ OutPlot2 = pairs.panels(my_corr_lig[1:(length(my_corr_lig)-1)]
#, alpha = .05 #, alpha = .05
#, points(pch = 19, col = c("#f8766d", "#00bfc4")) #, points(pch = 19, col = c("#f8766d", "#00bfc4"))
, cex = 3 , cex = 3
, cex.axis = 2.5 , cex.axis = 1.5
, cex.labels = 2.1 , cex.labels = 2.1
, cex.cor = 1 , cex.cor = 1
, smooth = F , smooth = F
@ -261,5 +301,4 @@ OutPlot2 = pairs.panels(my_corr_lig[1:(length(my_corr_lig)-1)]
print(OutPlot2) print(OutPlot2)
dev.off() dev.off()
####################################################### #######################################################
library(lattice)

103
scripts/plotting/corr_plots_style2.R → scripts/plotting/corr_plots_adjusted.R
View file

@ -14,7 +14,6 @@ getwd()
source("Header_TT.R") source("Header_TT.R")
require(cowplot) require(cowplot)
source("combining_dfs_plotting.R") source("combining_dfs_plotting.R")
source("my_pairs_panel.R") # with lower panel turned off
# should return the following dfs, directories and variables # should return the following dfs, directories and variables
# PS combined: # PS combined:
@ -56,12 +55,12 @@ cat(paste0("Variables imported:"
#plot_corr_plot_combined = paste0(plotdir,"/", corr_plot_combined) #plot_corr_plot_combined = paste0(plotdir,"/", corr_plot_combined)
# PS # PS
corr_ps_s2 = "corr_PS_style2.svg" corr_ps_adjusted = "corr_PS_adjusted.svg"
plot_corr_ps_s2 = paste0(plotdir,"/", corr_ps_s2) plot_corr_ps_adjusted = paste0(plotdir,"/", corr_ps)
# LIG # LIG
corr_lig_s2 = "corr_LIG_style2.svg" corr_lig_adjusted = "corr_LIG_adjusted.svg"
plot_corr_lig_s2 = paste0(plotdir,"/", corr_lig_s2) plot_corr_lig_adjusted = paste0(plotdir,"/", corr_lig)
#################################################################### ####################################################################
# end of loading libraries and functions # # end of loading libraries and functions #
@ -78,76 +77,33 @@ rm( merged_df2, merged_df2_comp, merged_df2_lig, merged_df2_comp_lig, my_df_u, m
# end of data extraction and cleaning for plots # # end of data extraction and cleaning for plots #
######################################################################## ########################################################################
#============================ #===========================
# adding foldx scaled values # Data for Correlation plots:PS
# scale data b/w -1 and 1 #===========================
#============================ table(df_ps$duet_outcome)
n = which(colnames(df_ps) == "ddg"); n
my_min = min(df_ps[,n]); my_min
my_max = max(df_ps[,n]); my_max
df_ps$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()
}
table(df_ps$foldx_outcome)
#======================
# adding log cols # adding log cols
#======================
df_ps$log10_or_mychisq = log10(df_ps$or_mychisq) df_ps$log10_or_mychisq = log10(df_ps$or_mychisq)
df_ps$neglog_pval_fisher = -log10(df_ps$pval_fisher) df_ps$neglog_pval_fisher = -log10(df_ps$pval_fisher)
df_ps$log10_or_kin = log10(df_ps$or_kin) df_ps$log10_or_kin = log10(df_ps$or_kin)
df_ps$neglog_pwald_kin = -log10(df_ps$pwald_kin) df_ps$neglog_pwald_kin = -log10(df_ps$pwald_kin)
#===========================
# Data for Correlation plots:PS
#===========================
# subset data to generate pairwise correlations # subset data to generate pairwise correlations
cols_to_select = c("duet_scaled" cols_to_select = c("duet_scaled"
, "foldx_scaled"
, "log10_or_mychisq" , "log10_or_mychisq"
, "neglog_pval_fisher" , "neglog_pval_fisher"
#, "or_kin"
#, "neglog_pwald_kin"
, "af" , "af"
, "duet_outcome" , "duet_outcome"
, drug) , drug)
corr_data_ps = df_ps[cols_to_select] corr_data_ps = df_ps[, cols_to_select]
dim(corr_data_ps) dim(corr_data_ps)
@ -157,11 +113,12 @@ dim(corr_data_ps)
# assign nice colnames (for display) # assign nice colnames (for display)
my_corr_colnames = c("DUET" my_corr_colnames = c("DUET"
, "Foldx"
, "Log(OR)" , "Log(OR)"
, "-Log(P)" , "-Log(P)"
#, "OR adjusted"
#, "-Log(P wald)"
, "AF" , "AF"
, "duet_outcome" , "duet_outcome"
@ -187,11 +144,10 @@ head(my_corr_ps)
# deep blue :#007d85 # deep blue :#007d85
# deep red: #ae301e # deep red: #ae301e
cat("Corr plot PS:", plot_corr_ps_s2) cat("Corr plot PS:", plot_corr_ps_adjusted)
svg(plot_corr_ps_s2, width = 15, height = 15) svg(plot_corr_ps_adjusted, width = 15, height = 15)
#OutPlot1 = pairs.panels([1:(length(my_corr_ps)-1)] OutPlot1 = pairs.panels(my_corr_ps[1:(length(my_corr_ps)-1)]
OutPlot1 = my_pp(my_corr_ps[1:(length(my_corr_ps)-1)]
, method = "spearman" # correlation method , method = "spearman" # correlation method
, hist.col = "grey" ##00AFBB , hist.col = "grey" ##00AFBB
, density = TRUE # show density plots , density = TRUE # show density plots
@ -206,11 +162,10 @@ OutPlot1 = my_pp(my_corr_ps[1:(length(my_corr_ps)-1)]
#, alpha = .05 #, alpha = .05
#, points(pch = 19, col = c("#f8766d", "#00bfc4")) #, points(pch = 19, col = c("#f8766d", "#00bfc4"))
, cex = 3 , cex = 3
, cex.axis = 1.5 , cex.axis = 2.5
, cex.labels = 2.1 , cex.labels = 2.1
, cex.cor = 1 , cex.cor = 1
, smooth = F , smooth = F
) )
print(OutPlot1) print(OutPlot1)
@ -235,6 +190,9 @@ cols_to_select = c("affinity_scaled"
, "log10_or_mychisq" , "log10_or_mychisq"
, "neglog_pval_fisher" , "neglog_pval_fisher"
#, "or_kin"
#, "neglog_pwald_kin"
, "af" , "af"
, "ligand_outcome" , "ligand_outcome"
@ -251,6 +209,9 @@ my_corr_colnames = c("Ligand Affinity"
, "Log(OR)" , "Log(OR)"
, "-Log(P)" , "-Log(P)"
#, "OR adjusted"
#, "-Log(P wald)"
, "AF" , "AF"
, "ligand_outcome" , "ligand_outcome"
@ -273,11 +234,10 @@ offset = 1
my_corr_lig = corr_data_lig[start:(end-offset)] my_corr_lig = corr_data_lig[start:(end-offset)]
head(my_corr_lig) head(my_corr_lig)
cat("Corr LIG plot:", plot_corr_lig_s2) cat("Corr LIG plot:", plot_corr_lig_adjusted)
svg(plot_corr_lig_s2, width = 15, height = 15) svg(plot_corr_lig_adjusted, width = 15, height = 15)
#OutPlot2 = pairs.panels(my_corr_lig[1:(length(my_corr_lig)-1)] OutPlot2 = pairs.panels(my_corr_lig[1:(length(my_corr_lig)-1)]
OutPlot2 = my_pp(my_corr_lig[1:(length(my_corr_lig)-1)]
, method = "spearman" # correlation method , method = "spearman" # correlation method
, hist.col = "grey" ##00AFBB , hist.col = "grey" ##00AFBB
, density = TRUE # show density plots , density = TRUE # show density plots
@ -292,7 +252,7 @@ OutPlot2 = my_pp(my_corr_lig[1:(length(my_corr_lig)-1)]
#, alpha = .05 #, alpha = .05
#, points(pch = 19, col = c("#f8766d", "#00bfc4")) #, points(pch = 19, col = c("#f8766d", "#00bfc4"))
, cex = 3 , cex = 3
, cex.axis = 1.5 , cex.axis = 2.5
, cex.labels = 2.1 , cex.labels = 2.1
, cex.cor = 1 , cex.cor = 1
, smooth = F , smooth = F
@ -301,4 +261,5 @@ OutPlot2 = my_pp(my_corr_lig[1:(length(my_corr_lig)-1)]
print(OutPlot2) print(OutPlot2)
dev.off() dev.off()
####################################################### #######################################################
library(lattice)