update tdg09 FDR -> LRT + plot

lib/convergence_detection.ml

@@ -99,7 +99,7 @@ let plot_merge_results ?t_choices ~plot_all_sites ~(res_by_tools:result list) ~t
         | `Identical_WAG _ -> "Identical_WAG01"
         | `Topological_LG _ -> "Topological_LG08"
         | `Topological_WAG _ -> "Topological_WAG01"
-        | `Tdg09 _ -> "Tdg09_1MinusFDR,Tdg09_prob_post"
+        | `Tdg09 _ -> "Tdg09_1MinusFDR,Tdg09_1MinusLRT,Tdg09_prob_post"
         | `Multinomial _ -> "Mutinomial_1MinusLRT"
         | `Msd _ -> "Msd_0.05_1MinusP"
       in
@@ -116,7 +116,7 @@ let plot_merge_results ?t_choices ~plot_all_sites ~(res_by_tools:result list) ~t
         | `Identical_WAG _ -> "Identical_WAG01:0.9"
         | `Topological_LG _ -> "Topological_LG08:0.9"
         | `Topological_WAG _ -> "Topological_WAG01:0.9"
-        | `Tdg09 _ -> "Tdg09_1MinusFDR:0.99,Tdg09_prob_post:0.99"
+        | `Tdg09 _ -> "Tdg09_1MinusFDR:0.99,Tdg09_prob_post:0.99,Tdg09_1MinusLRT:0.99"
         | `Multinomial _ -> "Mutinomial_1MinusLRT:0.95"
         | `Msd _ -> "Msd_0.05_1MinusP:0.95"
 

lib/post_analyses.ml

@@ -16,6 +16,7 @@ type t_choices = {
   t_choices_recall09: text_file workflow ;
   t_choices_plot: text_file workflow ;
   t_choices_condensed_plot: text_file workflow ;
+  t_choices_condensed_plot_png: text_file workflow ;
   rp_plot: text_file workflow ;
   tree_prefix: string;
 }
@@ -288,9 +289,10 @@ let get_t_choices ~tree_prefix ~(dataset_results_l: dataset_res list) : t_choice
     let t_choices_complete = t_choices_dir / selector ["out.complete.tsv"] in
     let t_choices_plot = t_choices_dir / selector ["out.pdf"] in
     let t_choices_condensed_plot = t_choices_dir / selector ["out_condensed.pdf"] in
+    let t_choices_condensed_plot_png = t_choices_dir / selector ["out_condensed.png"] in
     let rp_plot = t_choices_dir / selector ["out_condensed.recall_precision_ok.pdf"] in
     let tree_prefix = ha_PCOC.tree_prefix in
-    Some {t_choices_max; t_choices_recall09; t_choices_complete ; t_choices_plot; t_choices_condensed_plot; rp_plot; tree_prefix}
+    Some {t_choices_max; t_choices_recall09; t_choices_complete ; t_choices_plot; t_choices_condensed_plot;t_choices_condensed_plot_png;rp_plot; tree_prefix}
   | _ -> None
 
 
@@ -441,6 +443,7 @@ let repo_of_post_analyses_res ~prefix ~post_analyses_res =
          Repo.[
            item [prefix ^ ".t_choices.pdf"] w.t_choices_plot ;
            item [prefix ^ ".t_choices.condensed.pdf"] w.t_choices_condensed_plot ;
+           item [prefix ^ ".t_choices.condensed.png"] w.t_choices_condensed_plot_png ;
            item [prefix ^ ".recall_precision_ok.pdf"] w.rp_plot ;
          ]
        ] |> List.concat

lib/scripts/calc_t_per_meth.R

@@ -587,6 +587,7 @@ plot_out = function(df_out, df_d , df_recall_sup09_per_meth, meths = NULL, suffi
     limitsize = FALSE
     )
 
+    plot_value_JSD = NULL
     if ( all( c("P_JSD","P_ED", "C1", "C2", "entropy_C1","entropy_C2") %in% colnames(df_d))){
 
         print("plot value/distance")
@@ -673,23 +674,42 @@ plot_out = function(df_out, df_d , df_recall_sup09_per_meth, meths = NULL, suffi
 
 
     }
-    
-    plot = plot_grid(plot_recall_precision,plot_max_MCC,plot_value_JSD,plot_value_ED,plot_entropy_C2_C1,plot_entropy_C1,plot_entropy_C2,
-                     labels = c("A", "B","C","D","E","F"),
-                     rel_heights = c(length(unique(df_out$couple))*0.8,3,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8),
-                     nrow=7
-                     )
-    
-    #plot = plot_grid(plot_recall_precision,plot_recall_precision_papier,plot_max_MCC,plot_FPR,plot_FPR2,
-    #                labels = c("A", "A'", "B","B'","B''"),
-    #                rel_heights = c(length(unique(df_out$couple))*0.8,2,4,2,2),
-    #                nrow=5
+
+    #plot = plot_grid(plot_recall_precision,plot_max_MCC,plot_value_JSD,plot_value_ED,plot_entropy_C2_C1,plot_entropy_C1,plot_entropy_C2,
+    #                 labels = c("A", "B","C","D","E","F"),
+    #                 rel_heights = c(length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8,length(unique(df_out$couple))*0.8),
+    #                 nrow=7
     #                 )
+    plot = plot_grid(plot_recall_precision,plot_recall_precision_papier,plot_max_MCC,plot_FPR,plot_FPR2,
+                    labels = c("A", "A'", "B","B'","B''"),
+                    rel_heights = c(length(unique(df_out$couple))*0.8,2,4,2,2),
+                    nrow=5)
+
+    if (is.null(plot_value_JSD)) {
+        plot_png = plot_grid(plot_recall_precision,plot_max_MCC,
+                    labels = c("A", "B"),
+                    rel_heights = c(1,1),
+                    nrow=2
+                     )
+    } else {
+        plot_png = plot_grid(plot_recall_precision,plot_max_MCC,plot_value_JSD,
+                    labels = c("A", "B","C"),
+                    rel_heights = c(0.5,0.5,1),
+                    nrow=3
+                     )
+    }
 
     save_plot(paste0(opt$out,suffix,".pdf"),
               plot,
               ncol = 0.4* length(unique(df_out_melt$methode)),
-              nrow = length(unique(df_out$couple)) * 0.5  + 1 + 1 +1 +1,
+              nrow = length(unique(df_out$couple)) * 0.5  + 2 + 1 +1 +1,
+              base_aspect_ratio = 1,
+              limitsize = FALSE
+              )
+    save_plot(paste0(opt$out,suffix,".png"),
+              plot_png,
+              ncol = 0.4* length(unique(df_out_melt$methode)),
+              nrow = length(unique(df_out$couple)) * 1.5,
               base_aspect_ratio = 1,
               limitsize = FALSE
               )
@@ -698,7 +718,7 @@ plot_out = function(df_out, df_d , df_recall_sup09_per_meth, meths = NULL, suffi
 
 
 plot_out(df_out, df_d, df_recall_sup09_per_meth, suffix="")
-condensed_meths = c("PCOC","Diffsel_mean","Identical_LG08","Mutinomial_1MinusLRT","Tdg09_1MinusFDR","Msd_1MinusP","Msd_0.01_1MinusP","Msd_0.05_1MinusP","Msd_0.10_1MinusP","Msd_0.50_1MinusP","Msd_1.00_1MinusP","Topological_LG08")
+condensed_meths = c("PCOC","Diffsel_mean","Identical_LG08","Mutinomial_1MinusLRT","Tdg09_1MinusLRT","Msd_1MinusP","Msd_0.05_1MinusP","Topological_LG08")
 plot_out(df_out, df_d, df_recall_sup09_per_meth, meths=condensed_meths,  suffix="_condensed")
 
 

lib/scripts/parse_results_tdg09.py

@@ -43,6 +43,7 @@ Tdg09File.close()
 
 n_sites=int(searchlines[9].split()[1])
 fdr=["NA"] * (n_sites)
+lrt=["NA"] * (n_sites)
 prob_post=["NA"] * (n_sites)
 
 for i,line in enumerate(searchlines):
@@ -69,14 +70,18 @@ for i,line in enumerate(searchlines):
                     lnLconv   = lnLconv.strip()
                     lnLnoconv = lnLnoconv.strip()
                     FDR       = FDR.strip()
+                    LRT       = LRT.strip()
                     if Site and FDR != "NA":
                         FDR = float(FDR)
+                        LRT = float(LRT)
                         if FDR > 1:
                             FDR = 1
                         fdr[int(Site)-1]=1-FDR
+                        lrt[int(Site)-1]=1-LRT
                     if Site and FDR == "NA":
                         FDR = 1
                         fdr[int(Site)-1]=1-FDR
+                        lrt[int(Site)-1]=0
 
                     if Site and lnLconv != "NA" and lnLnoconv != "NA":
                         lnLconv = float(lnLconv)
@@ -90,11 +95,12 @@ for i,line in enumerate(searchlines):
 
 Sites = [i +1 for i in range(n_sites)]
 df_final = pd.DataFrame({'Sites': Sites,
+                  'Tdg09_1MinusLRT' : lrt,
                   'Tdg09_1MinusFDR' : fdr,
                   'Tdg09_prob_post' : prob_post})
 
 
-df_final = df_final[["Sites","Tdg09_1MinusFDR","Tdg09_prob_post"]]
+df_final = df_final[["Sites","Tdg09_1MinusFDR","Tdg09_1MinusLRT","Tdg09_prob_post"]]