fix to plot_grid

CHANGES

@@ -15,6 +15,8 @@ added neo module for neo-epitope prediction pipeline
 snap package now builds with some predictors built in
 fixed netmhciipan predictor
 default config file in home dir for paths
+changes default cutoff method for binders to use pre-defined quantiles for cutoffs
+simplified web interface
 
 ------
 0.2.0

epitopepredict/app.py

@@ -187,7 +187,8 @@ def get_summary(self, P, pb, seqs):
         x = pb.groupby('name').agg({'peptide':[base.first,np.size],
                                     P.scorekey:base.first}).reset_index()
         x.columns = [col[1]+'_'+col[0] if col[1]!='' else col[0] for col in x.columns.values]
-        x = x.rename(columns={'size_peptide':'binders','first_score':'max_score','first_peptide':'top_peptide'})
+        x = x.rename(columns={'size_peptide':'binders','first_%s' %P.scorekey:'max_score',
+                              'first_peptide':'top_peptide'})
         cl = analysis.find_clusters(pb)
         if len(cl)>0:
             cl = cl.groupby('name').size().rename('clusters').to_frame().reset_index()

epitopepredict/config.py

@@ -43,15 +43,14 @@
                 'mhc2_length': 11,
                 'n': 2, #number of alleles
                 'cutoff_method': 'default',
-                'cutoffs': 4, #percentile cutoff
+                'cutoffs': .95, #percentile cutoff
                 'sequence_file':'', #genbank/fasta file
                 'peptide_file':'', #plain text list of peptides
                 'path': 'results',
                 'overwrite': 'no',
                 'verbose':'no',
                 'names': '', #subset of protein names
                 'overwrite': 'no',
-                #'genome_analysis': 'no',
                 'cpus': 1,
                 'compression': '',
                 'fasta_header_sep': ' '}

epitopepredict/neo.py

@@ -50,12 +50,17 @@ def setup(self):
         elif self.mhc2_alleles[0] in base.mhc2_presets:
             self.mhc2_alleles = base.get_preset_alleles(self.mhc2_alleles[0])
 
-        if type(self.cutoffs) is int:
-            self.cutoffs = [self.cutoffs for p in self.predictors]
+        if type(self.cutoffs) is int or type(self.cutoffs) is float:
+            self.cutoffs = [self.cutoffs]
         else:
             self.cutoffs = [float(i) for i in self.cutoffs.split(',')]
-        self.names = self.names.split(',')
-        if self.names == ['']: self.names=None
+
+        if os.path.exists(self.names):
+            self.names = read_names(self.names)
+        elif self.names == '':
+            self.names=None
+        else:
+            self.names = self.names.split(',')
         if not os.path.exists(self.path) and self.path != '':
             os.mkdir(self.path)
         return True
@@ -77,8 +82,11 @@ def run(self):
         path = self.path
         overwrite = self.overwrite
         files = self.vcf_files
+        preds = self.predictors
         labels = self.get_file_labels(files)
         cutoffs = self.cutoffs
+        if len(cutoffs) < len(preds) :
+            cutoffs = [cutoffs[0] for p in preds]
 
         for f in labels:
             print (f)

epitopepredict/plotting.py

@@ -282,7 +282,12 @@ def bokeh_plot_grid(pred, name, width=None, palette='Blues', **kwargs):
         return
 
     cols = ['allele','pos','peptide',P.scorekey]
-    d = df[cols].copy()
+    #d = df[cols].copy()
+    b = P.get_binders(**kwargs)
+    d = P.data.copy()
+    #mark binders
+    mask = d.index.isin(b.index)
+    d['binder'] = mask
 
     l = base.get_length(df)
     grps = df.groupby('allele')
@@ -299,8 +304,7 @@ def bokeh_plot_grid(pred, name, width=None, palette='Blues', **kwargs):
     if P.name not in ['tepitope']:
         d['score1'] = d[val].apply( lambda x: 1-math.log(x, 50000))
         val='score1'
-        cut = .42
-    d[val] = d[val].clip(cut)
+    d[val][d.binder==False] = min(d[val])
 
     source = ColumnDataSource(d)
     colors = all_palettes[palette][7]
@@ -312,7 +316,7 @@ def bokeh_plot_grid(pred, name, width=None, palette='Blues', **kwargs):
     p.rect(x="pos", y="allele", width=1, height=1,
            source=source,
            fill_color={'field': val,'transform':mapper},
-           line_color='gray', line_width=.2)
+           line_color='gray', line_width=.5)
     p.select_one(HoverTool).tooltips = [
          ('allele', '@allele'),
          (P.scorekey, '@%s{1.11}' %P.scorekey),

epitopepredict/templates/global.html

@@ -17,14 +17,12 @@
 		<div class="floatleft">
 			{{ tables }}
 		</div>
-		{% end %}
 
 		<div class="floatright">
 		  {{ script }}
 		  {{ div }}
-
 		</div>
-
+		{% end %}
 	  </div>
 
 {% end %}

epitopepredict/templates/index.html

@@ -6,7 +6,7 @@
 		<table class="largetable">
 			<tr><th>
 			<a href="/global" class="largebtn">
-			<span>all results</span>
+			<span>summary</span>
 			</a> </th>
 			<td>
 			<p>Global view showing results for all protein sequences in a set of predictions.</p></td>