svm validation, tensorflow mp2d pa predictions, r results removed, cleanup

7 files changed, 83 insertions, 22 deletions

diff --git a/scripts/cv-tensorflow-confusion-matrix.rb b/scripts/cv-tensorflow-confusion-matrix.rb
index fba5c49..087d905 100755
--- a/scripts/cv-tensorflow-confusion-matrix.rb
+++ b/scripts/cv-tensorflow-confusion-matrix.rb
@@ -7,7 +7,7 @@ tn = 0
 fn = 0
 
 pred = CSV.read(ARGV[0],headers: true,:col_sep => ",")
-act = CSV.read(File.join("data","mutagenicity.csv"),headers: true,:col_sep => ",")
+act = CSV.read(File.join("data","training","mutagenicity.csv"),headers: true,:col_sep => ",")
 
 data = {}
 
diff --git a/scripts/mp2d-distances.rb b/scripts/mp2d-distances.rb
index 398504d..05313c2 100755
--- a/scripts/mp2d-distances.rb
+++ b/scripts/mp2d-distances.rb
@@ -17,7 +17,7 @@ File.readlines(File.join("..","lazar","models","mutagenicity","independent_varia
   end
 end
 
-File.readlines(File.join("pyrrolizidine-alkaloids","PA-smiles.csv")).each_with_index do |line,i|
+File.readlines(File.join("pyrrolizidine-alkaloids","lazar","pa-smiles.csv")).each_with_index do |line,i|
   if i > 0
     (id,smiles) = line.chomp.split(",")
     independent_variables << Compound.new(smiles).fingerprint
diff --git a/scripts/pa-fingerprints.rb b/scripts/pa-fingerprints.rb
new file mode 100755
index 0000000..344ba86
--- /dev/null
+++ b/scripts/pa-fingerprints.rb
@@ -0,0 +1,20 @@
+#!/usr/bin/env ruby
+require_relative "../../lazar/lib/lazar.rb"
+training_fingerprints = `sed -n '1p' data/mutagenicity-fingerprints.csv`.chomp.split(",")
+training_fingerprints.pop
+puts training_fingerprints.join(",")
+training_fingerprints.shift
+
+File.readlines(File.join("pyrrolizidine-alkaloids","lazar","pa-smiles.csv")).each_with_index do |line,i|
+  if i > 0
+    (id,smiles) = line.chomp.split(",")
+    c = Compound.new(smiles)
+    out = [c.smiles]
+    fp = c.fingerprint
+    training_fingerprints.each do |frag|
+      fp.include?(frag) ? out << 1 : out << 0
+    end
+    puts out.join(",")
+  end
+end
+
diff --git a/scripts/pa-groups.R b/scripts/pa-groups.R
new file mode 100755
index 0000000..005dd3d
--- /dev/null
+++ b/scripts/pa-groups.R
@@ -0,0 +1,23 @@
+#!/usr/bin/env Rscript
+library(ggplot2)
+data <- read.csv("tables/pa-table.csv")
+for (i in c(2:10)) {
+  name <- names(data)[i]
+  group <- data[data[i] == 1,c(15,19,20,21,22)]
+  freq <- 100*colSums(group,na.rm=TRUE)/colSums(!is.na(group))
+  plot <- ggplot(data.frame(freq),aes(x=c("lazar","LR-sgd","LR-scikit","NN","RF"),y=freq)) + geom_bar(stat="identity") + ylab("% mutagenic") + xlab(name) + ylim(c(0,100))
+  ggsave(paste("figures/",name,".png",sep=""))
+}
+#groups <- names(data)[c(2:10)]
+#labels <- data$Mutagenicity
+#data$Mutagenicity <- NULL
+#m <- as.matrix(data)
+#dist <- as.dist(m)
+#tsne <- Rtsne(dist,is_distance=T)
+#write.csv(tsne,"tsne.csv")
+#write.csv(data.frame(x = tsne$Y[,1], y = tsne$Y[,2]),"tsne.csv")
+#tsne_plot <- data.frame(x = tsne$Y[,1], y = tsne$Y[,2])
+#colors <- c("PA" = "#00BFC4", "mutagen" = "#F8766D", "non-mutagen" = "#7CAE00")
+#plot <- ggplot(tsne_plot)
+#plot + geom_point(aes(x=x, y=y, color = labels)) + xlab(element_blank()) + ylab(element_blank()) + theme(axis.ticks = element_blank(), axis.text = element_blank(), legend.title=element_blank()) + scale_color_manual(values = colors)
+#ggsave("figures/tsne-mp2d.png")
diff --git a/scripts/pa-table.rb b/scripts/pa-table.rb
index 8c1037e..1b8ecc8 100755
--- a/scripts/pa-table.rb
+++ b/scripts/pa-table.rb
@@ -2,7 +2,7 @@
 
 # red groups
 tab = []
-File.read("pyrrolizidine-alkaloids/R/PA.RF.outcome.csv").each_line do |l|
+File.read("data/pyrrolizidine-alkaloids/pa-groups.csv").each_line do |l|
   items = l.chomp.split(';')
   if items.first.empty?
     items[0] = "ID"
@@ -17,7 +17,7 @@ end
 tab[0] += ["CID","SMILES","Canonical SMILES","Measured","lazar-MP2D","lazar-MP2D-high-confidence","lazar-CDK","lazar-CDK-high-confidence"]
 
 i = 0
-File.read("pyrrolizidine-alkaloids/180920_PA_complete_SMILES.csv").each_line do |l|
+File.read("data/pyrrolizidine-alkaloids/180920_PA_complete_SMILES.csv").each_line do |l|
   if i > 0
     id,cid,name,smi = l.chomp.split(";")
     tab[i] += [cid,'"'+smi+'"']
@@ -26,7 +26,7 @@ File.read("pyrrolizidine-alkaloids/180920_PA_complete_SMILES.csv").each_line do
 end
 
 i = 0
-File.read("pyrrolizidine-alkaloids/lazar/pa-mp2d-predictions.csv").each_line do |l|
+File.read("pyrrolizidine-alkaloids/mp2d/lazar/pa-mp2d-predictions.csv").each_line do |l|
   if i > 0
     id,cansmi,exp,mut,p0,p1,max_sim,nn = l.chomp.split(",")
     max_sim.to_f < 0.5? hc = "F" : hc = "T"
@@ -36,8 +36,8 @@ File.read("pyrrolizidine-alkaloids/lazar/pa-mp2d-predictions.csv").each_line do
   i += 1
 end
 
-i=0
-File.read("pyrrolizidine-alkaloids/lazar/pa-cdk-predictions.csv").each_line do |l|
+i=1
+File.read("pyrrolizidine-alkaloids/cdk/lazar/pa-cdk-predictions.csv").each_line do |l|
   #if i > 0
     cansmi,exp,mut,p0,p1,max_sim,nn = l.chomp.split(",")
     max_sim.to_f < 0.5? hc = "F" : hc = "T"
@@ -47,6 +47,7 @@ File.read("pyrrolizidine-alkaloids/lazar/pa-cdk-predictions.csv").each_line do |
   i += 1
 end
 
+=begin
 Dir["pyrrolizidine-alkaloids/R/PA.*.outcome.csv"].each do |r|
   tab[0] << "R-"+r.sub('pyrrolizidine-alkaloids/R/PA.','').sub('.outcome.csv','')
   i = 0
@@ -63,9 +64,22 @@ Dir["pyrrolizidine-alkaloids/R/PA.*.outcome.csv"].each do |r|
     i += 1
   end
 end
+=end
 
-Dir["pyrrolizidine-alkaloids/tensorflow/pred.*.v5-ext-Padel-2D.csv"].each do |r|
-  tab[0] << "TF-"+r.sub('pyrrolizidine-alkaloids/tensorflow/pred.','').sub('.v5-ext-Padel-2D.csv','').sub("lr2","LR-scikit").sub("lr","LR-sgd").sub("rf","RF").sub("nn","NN")
+Dir["pyrrolizidine-alkaloids/cdk/tensorflow/*.csv"].each do |r|
+  tab[0] << "TF-"+r.sub('pyrrolizidine-alkaloids/tensorflow/pred.','').sub('.v5-ext-Padel-2D.csv','').sub("lr2","LR-scikit").sub("lr","LR-sgd").sub("rf","RF").sub("nn","NN").sub("svm","SVM")
+  i = 0
+  File.read(r).each_line do |l|
+    if i > 0
+      id,pred = l.chomp.split(",")
+      pred.to_f > 0.5 ? tab[i] << 1 : tab[i] << 0
+    end
+    i += 1
+  end
+end
+
+Dir["pyrrolizidine-alkaloids/cdk/tensorflow/*.csv"].each do |r|
+  tab[0] << "TF-"+r.sub('pyrrolizidine-alkaloids/tensorflow/pred.','').sub('.v5-ext-Padel-2D.csv','').sub("lr2","LR-scikit").sub("lr","LR-sgd").sub("rf","RF").sub("nn","NN").sub("svm","SVM")
   i = 0
   File.read(r).each_line do |l|
     if i > 0
diff --git a/scripts/pa-tex-table.rb b/scripts/pa-tex-table.rb
index 0fe1410..b163ab3 100755
--- a/scripts/pa-tex-table.rb
+++ b/scripts/pa-tex-table.rb
@@ -13,16 +13,16 @@ puts '
 \caption{Summary of pyrrolizidine alkaloid predictions: red: mutagen, green: non-mutagen, grey: no prediction, dark red/green: low confidence; 1: Retronecine, 2: Otonecine, 3: Platynecine, 4: N-oxide, 5: Dehydropyrrolizidine, 6:Tertiary PA, 7: Macrocyclic-diester, 8: Monoester, 9: Diester} \\\\
 \label{tab:pa}
 
-1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & Measured & MP2D & CDK & DL & RF & SVM & LR-sgd & LR-scikit & NN & RF \\kill % needed as guide for multicolumn
-\multicolumn{9}{c}{PA Group} & & \multicolumn{2}{c}{lazar} & \multicolumn{3}{c}{R} & \multicolumn{4}{c}{Tensorflow}\\\\
+1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & Measured & MP2D & CDK & DL & SVM & LR-sgd & LR-scikit & NN & RF \\kill % needed as guide for multicolumn
+\multicolumn{9}{c}{PA Group} & & \multicolumn{2}{c}{lazar} & \multicolumn{4}{c}{Tensorflow}\\\\
 
-1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & Measured & MP2D & CDK & DL & RF & SVM & LR-sgd & LR-scikit & NN & RF \\\\
+1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & Measured & MP2D & CDK & LR-sgd & LR-scikit & NN & RF \\\\
 \hline
 \renewcommand{\arraystretch}{0.075}
 '
 File.read(ARGV[0]).each_line do |l|
   unless l.match("SMILES")
-    id,r,o,p,n,de,t,ma,mo,di,cid,smi,cansmi,exp,lazar_MP2D,lazar_MP2D_high_confidence,lazar_CDK,lazar_CDK_high_confidence,r_DL,r_RF,r_SVM,tf_lr_sgd,tf_lr_scikit,tf_NN,tf_RF = l.chomp.split(",")
+    id,r,o,p,n,de,t,ma,mo,di,cid,smi,cansmi,exp,lazar_MP2D,lazar_MP2D_high_confidence,lazar_CDK,lazar_CDK_high_confidence,tf_lr_sgd,tf_lr_scikit,tf_NN,tf_RF = l.chomp.split(",")
     row = [r,o,p,n,de,t,ma,mo,di].collect{|group| group == "1" ? '\cellcolor{black}' : '\cellcolor{white}'}.join(' & ')
     if exp == "1"
       row += ' & \cellcolor{red}'
@@ -49,7 +49,7 @@ File.read(ARGV[0]).each_line do |l|
     else 
       row += ' & \cellcolor{grey}'
     end
-    [r_DL,r_RF,r_SVM,tf_lr_sgd,tf_lr_scikit,tf_NN,tf_RF].each do |mut|
+    [tf_lr_sgd,tf_lr_scikit,tf_NN,tf_RF].each do |mut|
       mut == "1" ? row += ' & \cellcolor{red}' : row += ' & \cellcolor{green}'
     end
     puts row + ' \\\\'
diff --git a/scripts/summary2table.rb b/scripts/summary2table.rb
index 557dbd4..8bc323c 100755
--- a/scripts/summary2table.rb
+++ b/scripts/summary2table.rb
@@ -5,21 +5,25 @@ rows = {:acc => "Accuracy", :tpr => "True positive rate/Sensitivity", :tnr => "T
 data = YAML.load_file("10-fold-crossvalidations/summary.yaml")[:cv]
 
 case ARGV[0]
-when "R"
-  header = ["RF","SVM","DL"]
-  keys = header.collect{|h| "R-"+h}
 when "tensorflow"
-  header = ["RF","LR-sgd","LR-scikit","NN"]
-  keys = ["rf","lr","lr2","nn"].collect{|n| "tensorflow-"+n+".v3"}
+  header = ["MP2D-RF","MP2D-LR-sgd","MP2D-LR-scikit","MP2D-NN","MP2D-SVM","CDK-RF","CDK-LR-sgd","CDK-LR-scikit","CDK-NN","CDK-SVM"]
+  desc = ["mp2d","cdk"]
+  algos = ["rf","lr","lr2","nn","svm"]
+  keys = []
+  desc.each do |d| 
+    algos.each do |a|
+      keys << "tensorflow-"+a+"-"+d
+    end
+  end
 when "lazar"
   header = ["MP2D", "CDK"]
-  mp2dkeys = ["lazar-all","lazar-high-confidence"]
-  padelkeys = ["lazar-padel-all","lazar-padel-high-confidence"]
+  mp2dkeys = ["lazar-mp2d-all","lazar-mp2d-high-confidence"]
+  cdkkeys = ["lazar-cdk-all","lazar-cdk-high-confidence"]
   puts ","+header.join(",")
   rows.each do |short,long|
     print long+","
     print mp2dkeys.collect{|k| data[k][short]}.join("/")+","
-    puts padelkeys.collect{|k| data[k][short]}.join("/")
+    puts cdkkeys.collect{|k| data[k][short]}.join("/")
   end
   exit
 end