Fixing redundant column in LinearCount output

dieterich-lab · Jan 21, 2018 · 11fda2c · 11fda2c
1 parent 9b6470c
commit 11fda2c
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 3 deletions.
diff --git a/DCC/CombineCounts.py b/DCC/CombineCounts.py
@@ -143,8 +143,7 @@ def combine(self, Alist, col=7, circ=True):
                     res.setdefault(line_nr, ['\t'.join(line_split[:3]+[line_split[5]])]).append(line_split[col - 1])
 
                 else:  # input are host gene counts
-                    res.setdefault(line_nr, ['\t'.join(line_split[:3]+[line_split[5]])]).append(line_split[col - 1])
-
+                    res.setdefault(line_nr, ['\t'.join(line_split[:3])]).append(line_split[col - 1])
         return res
 
     def writeouput(self, output, res, samplelist=None, header=False):
@@ -155,3 +154,12 @@ def writeouput(self, output, res, samplelist=None, header=False):
         for line_nr in sorted(res):
             outfile.write("%s\n" % '\t'.join(res[line_nr]))
         outfile.close()
+
+    def writeouput_linear(self, output, res, samplelist=None, header=False):
+        # Sample list is a string with sample names seperated by \t.
+        outfile = open(output, 'w')
+        if header:
+            outfile.write('Chr\tStart\tEnd\t' + samplelist + '\n')
+        for line_nr in sorted(res):
+            outfile.write("%s\n" % '\t'.join(res[line_nr]))
+        outfile.close()
diff --git a/DCC/main.py b/DCC/main.py
@@ -440,7 +440,7 @@ def main():
 
             # Combine all to a individual sample host gene count to a single table
             res = cm.combine(linearfiles, col=6, circ=False)
-            cm.writeouput(output_linear_counts, res, samplelist, header=True)
+            cm.writeouput_linear(output_linear_counts, res, samplelist, header=True)
             logging.info("Finished combine individual linear gene expression counts")
 
             if not options.temp: