oops!

scripts/plotting/plotting_thesis/gg_pairs_all.R

@@ -0,0 +1,144 @@
+source("~/git/LSHTM_analysis/config/embb.R")
+source("~/git/LSHTM_analysis/scripts/plotting/plotting_colnames.R")
+source("~/git/LSHTM_analysis/scripts/plotting/get_plotting_dfs.R")
+
+my_gg_pairs=function(plot_df){
+  ggpairs(plot_df,
+          columns = 1:(ncol(plot_df)-1),
+          upper = list(
+            continuous = wrap('cor',
+                              method = "spearman",
+                              title="ρ",
+                              digits=2,
+                              justify_labels = "left",
+                              title_args=c(colour="black")
+            )
+          ),
+          lower = list(
+            continuous = wrap("points",
+                              alpha = 0.7,
+                              size=0.5),
+            combo     = wrap("dot",
+                             alpha = 0.7,
+                             size=0.5)
+          ),
+          aes(colour = factor(ifelse(plot_df$dst_mode==0,
+                                     "S",
+                                     "R") ),
+              alpha = 0.5),
+          title="Stability") +
+    
+    scale_colour_manual(values = c("red", "blue")) +
+    scale_fill_manual(values = c("red", "blue")) +
+    theme(text = element_text(size=12,
+                              face="bold") )
+}
+
+DistCutOff = 10
+
+merged_df3 = as.data.frame(merged_df3)
+corr_plotdf = corr_data_extract(merged_df3
+                                , gene = gene
+                                , drug = drug
+                                , extract_scaled_cols = F)
+
+aff_dist_cols = colnames(corr_plotdf)[grep("Dist", colnames(corr_plotdf))]
+static_cols =  c("Log10(MAF)"
+                 , "Log10(OR)")
+
+corr_ps_colnames = c(static_cols
+                     , "DUET"
+                     , "FoldX"
+                     , "DeepDDG"
+                     , "Dynamut2"
+                     , aff_dist_cols
+                     , "dst_mode")
+corr_df_ps = corr_plotdf[, corr_ps_colnames]
+complete_obs_ps = nrow(corr_df_ps) - sum(is.na(corr_df_ps$`Log(OR)`))
+color_coln = which(colnames(corr_df_ps) == "dst_mode")
+corr_end = color_coln-1
+
+# Plot #1
+plot_corr_df_ps = my_gg_pairs(corr_df_ps)
+
+
+corr_conservation_cols = c( static_cols
+                            , "ConSurf"
+                            , "SNAP2"
+                            , "PROVEAN"
+                            , aff_dist_cols
+                            , "dst_mode"
+)
+
+corr_df_cons = corr_plotdf[, corr_conservation_cols]
+complete_obs_cons = nrow(corr_df_cons) - sum(is.na(corr_df_cons$`Log(OR)`))
+color_coln = which(colnames(corr_df_cons) == "dst_mode")
+corr_end = color_coln-1
+
+# Plot #2
+
+#my_gg_pairs(corr_df_cons)
+plot_corr_df_cons = my_gg_pairs(corr_df_cons)
+
+
+corr_df_lig = corr_plotdf[corr_plotdf["Lig-Dist"]<DistCutOff,]
+corr_lig_colnames = c(static_cols
+                      , "mCSM-lig"
+                      , "mmCSM-lig"
+                      , "dst_mode")
+
+corr_df_lig = corr_plotdf[, corr_lig_colnames]
+
+complete_obs_lig = nrow(corr_df_lig) - sum(is.na(corr_df_lig$`Log(OR)`))
+color_coln = which(colnames(corr_df_lig) == "dst_mode")
+corr_end = color_coln-1
+# Plot #3
+
+#my_gg_pairs(corr_df_lig)
+plot_corr_df_lig = my_gg_pairs(corr_df_lig)
+
+corr_df_ppi2 = corr_plotdf[corr_plotdf["PPI-Dist"]<DistCutOff,]
+corr_ppi2_colnames = c(static_cols
+                       , "mCSM-PPI2"
+                       , "dst_mode"
+)
+corr_df_ppi2 = corr_plotdf[, corr_ppi2_colnames]
+complete_obs_ppi2 = nrow(corr_df_ppi2) - sum(is.na(corr_df_ppi2$`Log(OR)`))
+color_coln = which(colnames(corr_df_ppi2) == "dst_mode")
+corr_end = color_coln-1
+
+# NOTE: DELETE LOG OR FROM CORRELATION PLOTS!!!!!
+# NOTE: ALSO MAYBE DELETE DISTANCES AS WELL
+# NOTE: http://ggobi.github.io/ggally/reference/ggally_cor.html
+
+# "***" if the p-value is < 0.001
+# "**" if the p-value is < 0.01
+# "*" if the p-value is < 0.05
+# "." if the p-value is < 0.10
+# "" otherwise
+# 
+
+# Plot #4
+#my_gg_pairs(corr_df_ppi2)
+plot_corr_df_ppi2 = my_gg_pairs(corr_df_ppi2)
+
+
+# corr_df_na = corr_df_na[corr_df_na["NA-Dist"]<DistCutOff,]
+# corr_na_colnames = c(static_cols
+#                      , "mCSM-NA"
+#                      , "dst_mode"
+# )
+# 
+# corr_df_na = corr_plotdf[, corr_na_colnames]
+# complete_obs_na = nrow(corr_df_na) - sum(is.na(corr_df_na$`Log(OR)`))
+# color_coln = which(colnames(corr_df_na) == "dst_mode")
+# corr_end = color_coln-1
+# 
+# # Plot #5
+# #my_gg_pairs(corr_df_na)
+# plot_corr_df_na = my_gg_pairs(corr_df_na) 
+
+
+cowplot::plot_grid(ggmatrix_gtable(plot_corr_df_ps),ggmatrix_gtable(plot_corr_df_cons),
+                   ggmatrix_gtable(plot_corr_df_lig),ggmatrix_gtable(plot_corr_df_ppi2),
+                   nrow=2, ncol=2, rel_heights = 7,7,3,3)