reworked make_BN_from_carnaval

src/BayesPairing.py

@@ -9,7 +9,7 @@ from operator import itemgetter
 import ast
 import math
 import time
-
+import make_BN_from_carnaval as make_BN
 
 
 def find_loops(struct):
@@ -55,13 +55,14 @@ def parse_sequence(seq,modules,ss,dataset,left_out,sm=0.25, mc=0,k=5000):
     scores = {}
     for mod in modules:
         print("building Bayes Net for module : "+str(mod))
-        this_graph = graphs[mod][0]
        # print(this_graph.nodes())
-        os.chdir("../src")
-        call = str("python make_BN_from_carnaval.py"+" "+str(mod) +" "+ dataset +" " + str(left_out))
-        subprocess.call(call,shell=True)
-        os.chdir("../models")
+        #os.chdir("../src")
+        #call = str("python make_BN_from_carnaval.py"+" "+str(mod) +" "+ dataset +" " + str(left_out))
+        #subprocess.call(call,shell=True)
+        #os.chdir("../models")
+        make_BN.call_makeBN(mod,dataset,left_out)
         motif =  pickle.load( open( str(dataset+"_BN_"+str(mod)), "rb" ) )
+        this_graph = graphs[mod][0]
         constraints = testSS.get_constraints_from_BN(this_graph)
         pickle_name = dataset + "_samp"+str(mod)+".cPickle"
         if left_out != "NONE":

src/carnaval_to_rmdetect.py

@@ -5,8 +5,8 @@ from Bio.Align.Applications import ClustalwCommandline
 from Bio import AlignIO
 from Bio import SeqIO
 aligned_modulegraphs = pickle.load(open("../models/all_21+_rna3dmotif_106_aligned_modulegraphs.cPickle", 'rb'))
-PDB_positions = pickle.load(open("../models/all_21+_rna3dmotif_106__PDB_positions.cPickle", 'rb'))
-PDBs = pickle.load(open("../models/all_21+_rna3dmotif_106__PDB_names.cPickle", 'rb'))
+PDB_positions = pickle.load(open("../models/all_21+_rna3dmotif_106_PDB_positions.cPickle", 'rb'))
+PDBs = pickle.load(open("../models/all_21+_rna3dmotif_106_PDB_names.cPickle", 'rb'))
 graphs = pickle.load(open("../models/all_21+_rna3dmotif_106_one_of_each_graph.cPickle", 'rb'))
 module = 1
 aln = aligned_modulegraphs[module]

src/make_BN_from_carnaval.py

@@ -124,6 +124,65 @@ def make_graph_from_carnaval(g, aln_dict):
     motif.add_edges_from(edges_3d)
 
     return motif
+
+def call_makeBN(mod,dataset,left_out):
+    current_ID = mod
+    excluded = left_out
+
+    bayesname = dataset+"_BN_"+str(current_ID)
+    os.chdir("../models")
+    if excluded != "NONE":
+        removecall = "rm "+ bayesname
+        subprocess.call(removecall, shell=True)
+    #if not os.path.isfile(bayesname), changed for cross-validation; always remake the model
+    if not os.path.isfile(bayesname):
+         aln_list = pickle.load(open(dataset  + "_aligned_modulegraphs.cPickle",'rb'))
+         g_list = pickle.load(open(dataset + "_one_of_each_graph.cPickle",'rb'))
+         pdb_names  = pickle.load(open(dataset + "_PDB_names.cPickle",'rb'))
+         #extra_seqs  = pickle.load(open(dataset + "_rfam.cPickle",'rb'))[current_ID]
+         extra_seqs=""
+         if excluded!="NONE":
+             excluded_indexes = []
+             for pdb in range(len(pdb_names)):
+                if excluded in pdb_names[pdb]:
+                    excluded_indexes.append(pdb)
+
+             g = g_list[current_ID]
+             for k in reversed(excluded_indexes):
+                 g.pop(k)
+             aln = aln_list[current_ID]
+             for i in aln:
+                for k in reversed(excluded_indexes):
+                    i.pop(k)
+             alignment = get_alignment(g,aln,extra_seqs)
+         else:
+             g = g_list[current_ID]
+             aln = aln_list[current_ID]
+             alignment = get_alignment(g, aln,extra_seqs)
+         #print("ALIGNMENT")
+         #print(alignment)
+         addresses = get_addresses(aln)
+         #print("ADDRESSES")
+         #print(addresses)
+         seqs = get_seqs(g,aln,addresses)
+         #seqs = ""
+         #print("SEQUENCES")
+         #print(seqs)
+         aln_dict = (alignment,addresses,seqs)
+         #print(aln_dict)
+         motif = make_graph_from_carnaval(g, aln_dict)
+         pwm = BN_builder.build_pwm(motif.nodes(),aln_dict)
+         #print("PWM")
+         #print(pwm)
+         bob  = BN_builder.build_BN(motif,pwm,aln_dict)
+         pickle.dump(bob,open(bayesname,'wb'))
+         pdb_list = get_pdb_list(pdb_names[current_ID])
+         #seqs = get_pdb_seqs(pdb_list)
+         pickle.dump(seqs,open("seqs_"+str(bayesname),'wb'))
+    else:
+         print("Bayes Net exist, continuing")
+
+
 if __name__ == "__main__":
     print("ARGUMENTS")
     print(sys.argv)