Implement #709, plotCorrelation scatterplot with X and Y axis ranges.

deeptools/correlation.py

@@ -370,7 +370,7 @@ def plot_correlation(self, plot_filename, plot_title='', vmax=None,
         fig.savefig(plot_filename, format=image_format)
         plt.close()
 
-    def plotly_scatter(self, plot_filename, corr_matrix, plot_title='', minVal=None, maxVal=None):
+    def plotly_scatter(self, plot_filename, corr_matrix, plot_title='', minXVal=None, maxXVal=None, minYVal=None, maxYVal=None):
         """Make the scatter plot of a matrix with plotly"""
         n = self.matrix.shape[1]
         self.matrix = self.matrix
@@ -393,13 +393,13 @@ def plotly_scatter(self, plot_filename, corr_matrix, plot_title='', minVal=None,
             domain = [base, base + domainWidth]
             if x > 0:
                 base = 1 - base
-                fig['layout']['xaxis{}'.format(x + 1)] = dict(domain=domain, range=[minVal, maxVal], anchor='free', position=base)
+                fig['layout']['xaxis{}'.format(x + 1)] = dict(domain=domain, range=[minXVal, maxXVal], anchor='free', position=base)
             for y in range(0, n):
                 yanchor = 'y{}'.format(y + 1)
                 if x == 1:
                     base = 1 - y * domainWidth
                     domain = [base - domainWidth, base]
-                    fig['layout']['yaxis{}'.format(y + 1)] = dict(domain=domain, range=[minVal, maxVal], side='right', anchor='free', position=1.0)
+                    fig['layout']['yaxis{}'.format(y + 1)] = dict(domain=domain, range=[minYVal, maxYVal], side='right', anchor='free', position=1.0)
 
                 if x > y:
                     vector1 = self.matrix[:, x]
@@ -425,7 +425,7 @@ def plotly_scatter(self, plot_filename, corr_matrix, plot_title='', minVal=None,
 
         offline.plot(fig, filename=plot_filename, auto_open=False)
 
-    def plot_scatter(self, plot_filename, plot_title='', image_format=None, log1p=False, maxRange=None):
+    def plot_scatter(self, plot_filename, plot_title='', image_format=None, log1p=False, xRange=None, yRange=None):
         """
         Plot the scatter plots of a matrix
         in which each row is a sample
@@ -440,18 +440,28 @@ def plot_scatter(self, plot_filename, plot_title='', image_format=None, log1p=Fa
         plt.suptitle(plot_title)
         if log1p is True:
             self.matrix = np.log1p(self.matrix)
-        min_value = self.matrix.min()
-        max_value = self.matrix.max()
-        if maxRange is not None:
-            max_value = maxRange
-        if (min_value % 2 == 0 and max_value % 2 == 0) or \
-                (min_value % 1 == 0 and max_value % 2 == 1):
+        min_xvalue = self.matrix.min()
+        max_xvalue = self.matrix.max()
+        min_yvalue = min_xvalue
+        max_yvalue = max_xvalue
+        if xRange is not None:
+            min_xvalue = xRange[0]
+            max_xvalue = xRange[1]
+        if yRange is not None:
+            min_yvalue = yRange[0]
+            max_yvalue = yRange[1]
+        if (min_xvalue % 2 == 0 and max_xvalue % 2 == 0) or \
+                (min_xvalue % 1 == 0 and max_xvalue % 2 == 1):
             # make one value odd and the other even
-            max_value += 1
+            max_xvalue += 1
+        if (min_yvalue % 2 == 0 and max_yvalue % 2 == 0) or \
+                (min_yvalue % 1 == 0 and max_yvalue % 2 == 1):
+            # make one value odd and the other even
+            max_yvalue += 1
 
         # plotly output
         if image_format == 'plotly':
-            self.plotly_scatter(plot_filename, corr_matrix, plot_title=plot_title, minVal=min_value, maxVal=max_value)
+            self.plotly_scatter(plot_filename, corr_matrix, plot_title=plot_title, minXVal=min_xvalue, maxXVal=max_xvalue, minYVal=min_yvalue, maxYVal=max_yvalue)
             return
 
         rows, cols = np.triu_indices(num_samples)
@@ -518,12 +528,14 @@ def plot_scatter(self, plot_filename, plot_title='', image_format=None, log1p=Fa
 
             ax.hist2d(vector1, vector2, bins=200, cmin=0.1)
 
-            if maxRange is not None:
-                ax.set_xlim(min_value, min(maxRange, ax.get_xlim()[1]))
-                ax.set_ylim(min_value, min(maxRange, ax.get_ylim()[1]))
+            if xRange is not None:
+                ax.set_xlim(xRange)
+            else:
+                ax.set_xlim(min_xvalue, ax.get_xlim()[1])
+            if yRange is not None:
+                ax.set_ylim(min_yvalue, min(yRange, ax.get_ylim()[1]))
             else:
-                ax.set_xlim(min_value, ax.get_xlim()[1])
-                ax.set_ylim(min_value, ax.get_ylim()[1])
+                ax.set_ylim(min_yvalue, ax.get_ylim()[1])
 
         plt.savefig(plot_filename, format=image_format)
         plt.close()

deeptools/plotCorrelation.py

@@ -133,9 +133,16 @@ def scatterplot_options():
     parser = argparse.ArgumentParser(add_help=False)
     scatter_opts = parser.add_argument_group('Scatter plot options')
 
-    scatter_opts.add_argument('--maxRange',
-                              help='Maximum (integer) value for the X and Y axes. The default scales these such that the full range of dots is displayed. If you specify --log1p, please ensure that this value is within the range it will produce.',
+    scatter_opts.add_argument('--xRange',
+                              help='The X axis range. The default scales these such that the full range of dots is displayed.',
                               type=int,
+                              nargs=2,
+                              default=None)
+
+    scatter_opts.add_argument('--yRange',
+                              help='The Y axis range. The default scales these such that the full range of dots is displayed.',
+                              type=int,
+                              nargs=2,
                               default=None)
 
     scatter_opts.add_argument('--log1p',
@@ -229,7 +236,8 @@ def main(args=None):
             corr.plot_scatter(args.plotFile,
                               plot_title=args.plotTitle,
                               image_format=args.plotFileFormat,
-                              maxRange=args.maxRange,
+                              xRange=args.xRange,
+                              yRange=args.yRange,
                               log1p=args.log1p)
         else:
             corr.plot_correlation(args.plotFile,

galaxy/wrapper/plotCorrelation.xml

@@ -18,8 +18,11 @@
                 --plotHeight $plotting_type.plotHeight
             #else:
                 --plotTitle '$plotting_type.plotTitle'
-                #if str($plotting_type.maxRange) != '':
-                    --maxRange '$plotting_type.maxRange'
+                #if str($plotting_type.xRange) != '':
+                    --xRange '$plotting_type.xRange'
+                #end if
+                #if str($plotting_type.yRange) != '':
+                    --yRange '$plotting_type.yRange'
                 #end if
                 $plotting_type.log1p
             #end if
@@ -47,9 +50,12 @@
             </when>
             <when value="scatterplot">
                 <expand macro="plotTitle" />
-                <param argument="--maxRange" type="integer" value="" optional="true"
-                    label="Maximum Range"
-                    help="Maximum (integer) value for the X and Y axes. The default scales these such that the full range of dots is displayed. If you specify --log1p, please ensure that this value is within the range it will produce." />
+                <param argument="--xRange" value="" optional="true"
+                    label="X axis range"
+                    help="X axis range, the default scales these such that the full range of dots is displayed. The input should be two integers separated by a space." />
+                <param argument="--yRange" value="" optional="true"
+                    label="Y axis range"
+                    help="Y axis range, the default scales these such that the full range of dots is displayed. The input should be two integers separated by a space." />
                 <param argument="--log1p" type="boolean" truevalue="--log1p" falsevalue=""
                     label="Log transform"
                     help="Plot the natural log of the scatter plot after adding 1. Note that this is ONLY for plotting, the correlation is unaffected." />