outfactored bcmembers workflow

* bcmembers.py * fixed reference to background_df()
baliga-lab · Aug 30, 2024 · c4260d5 · c4260d5
1 parent 0a595ad
commit c4260d5
Show file tree

Hide file tree

Showing 3 changed files with 44 additions and 38 deletions.
diff --git a/bin/miner3-bcmembers b/bin/miner3-bcmembers
@@ -11,6 +11,7 @@ import matplotlib
 matplotlib.use('Agg')
 from miner import miner, util
 from miner import GIT_SHA, __version__ as pkg_version
+from miner import bcmembers
 
 
 DESCRIPTION = """miner3-bcmembers - MINER compute bicluster membership inference.
@@ -45,39 +46,5 @@ if __name__ == '__main__':
 
     with open(args.regulons) as infile:
         regulon_modules = json.load(infile)
+    bcmembers.bcmembers(exp_data, regulon_modules, args.outdir)
 
-    bkgd = miner.backgroundDf(exp_data)
-    overexpressed_members = miner.biclusterMembershipDictionary(regulon_modules,
-                                                                bkgd, label=2, p=0.05)
-    underexpressed_members = miner.biclusterMembershipDictionary(regulon_modules,
-                                                                 bkgd, label=0, p=0.05)
-    dysregulated_members = miner.biclusterMembershipDictionary(regulon_modules,
-                                                               bkgd, label="excluded")
-    coherent_members = miner.biclusterMembershipDictionary(regulon_modules,
-                                                           bkgd, label="included")
-
-    # write the overexpressed/underexpressed members as JSON, tools later in the pipeline can
-    # easier access them
-    with open(os.path.join(args.outdir, 'overExpressedMembers.json'), 'w') as out:
-        json.dump(overexpressed_members, out)
-    with open(os.path.join(args.outdir, 'underExpressedMembers.json'), 'w') as out:
-        json.dump(underexpressed_members, out)
-
-    overexpressed_members_matrix = miner.membershipToIncidence(overexpressed_members,
-                                                               exp_data)
-    overexpressed_members_matrix.to_csv(os.path.join(args.outdir,
-                                                     "overExpressedMembers.csv"))
-
-    underexpressed_members_matrix = miner.membershipToIncidence(underexpressed_members,
-                                                                exp_data)
-    underexpressed_members_matrix.to_csv(os.path.join(args.outdir,
-                                                      "underExpressedMembers.csv"))
-
-    dysregulated_members_matrix = miner.membershipToIncidence(dysregulated_members,
-                                                              exp_data)
-    dysregulated_members_matrix.to_csv(os.path.join(args.outdir, "dysregulatedMembers.csv"))
-
-    coherent_members_matrix = miner.membershipToIncidence(coherent_members,
-                                                          exp_data)
-    coherent_members_matrix.to_csv(os.path.join(args.outdir,
-                                                "coherentMembers.csv"))
diff --git a/miner/bcmembers.py b/miner/bcmembers.py
@@ -0,0 +1,42 @@
+"""bcmembers.py - Module to compute bicluster memberships"""
+
+import os
+import json
+from miner import miner
+
+def bcmembers(exp_data, regulon_modules, outdir):
+    bkgd = miner.background_df(exp_data)
+    overexpressed_members = miner.biclusterMembershipDictionary(regulon_modules,
+                                                                bkgd, label=2, p=0.05)
+    underexpressed_members = miner.biclusterMembershipDictionary(regulon_modules,
+                                                                 bkgd, label=0, p=0.05)
+    dysregulated_members = miner.biclusterMembershipDictionary(regulon_modules,
+                                                               bkgd, label="excluded")
+    coherent_members = miner.biclusterMembershipDictionary(regulon_modules,
+                                                           bkgd, label="included")
+
+    # write the overexpressed/underexpressed members as JSON, tools later in the pipeline can
+    # easier access them
+    with open(os.path.join(outdir, 'overExpressedMembers.json'), 'w') as out:
+        json.dump(overexpressed_members, out)
+    with open(os.path.join(outdir, 'underExpressedMembers.json'), 'w') as out:
+        json.dump(underexpressed_members, out)
+
+    overexpressed_members_matrix = miner.membershipToIncidence(overexpressed_members,
+                                                               exp_data)
+    overexpressed_members_matrix.to_csv(os.path.join(outdir,
+                                                     "overExpressedMembers.csv"))
+
+    underexpressed_members_matrix = miner.membershipToIncidence(underexpressed_members,
+                                                                exp_data)
+    underexpressed_members_matrix.to_csv(os.path.join(outdir,
+                                                      "underExpressedMembers.csv"))
+
+    dysregulated_members_matrix = miner.membershipToIncidence(dysregulated_members,
+                                                              exp_data)
+    dysregulated_members_matrix.to_csv(os.path.join(outdir, "dysregulatedMembers.csv"))
+
+    coherent_members_matrix = miner.membershipToIncidence(coherent_members,
+                                                          exp_data)
+    coherent_members_matrix.to_csv(os.path.join(outdir,
+                                                "coherentMembers.csv"))
diff --git a/miner/miner.py b/miner/miner.py
@@ -1142,7 +1142,6 @@ def biclusterMembershipDictionary(revisedClusters,background,label=2,p=0.05):
                 members[key] = []
                 continue
             members[key] = list(background.columns[nonMembers])
-        print("done!")
         return members
 
     if label == "included":
@@ -1158,7 +1157,6 @@ def biclusterMembershipDictionary(revisedClusters,background,label=2,p=0.05):
                 members[key] = []
                 continue
             members[key] = list(background.columns[included])
-        print("done!")
         return members
 
     members = {}
@@ -1174,7 +1172,6 @@ def biclusterMembershipDictionary(revisedClusters,background,label=2,p=0.05):
             members[key] = []
             continue
         members[key] = list(background.columns[overExpMembers])
-    print("done!")
     return members
 
 def membershipToIncidence(membershipDictionary,expressionData):