improved accuracy on new sequences, alleviated overfitting to sequence distance

bayespairing/src/bayes_to_seqpy.py

@@ -586,6 +586,10 @@ def seq_to_struct(
     g_distance=[]
     h_distance = []
     j_distance = []
+    i_distance = []
+    k_distance = []
+    l_distance = []
+    m_distance = []
     for junc in component_distance:
         a=junc[0]
         b=junc[1]
@@ -595,6 +599,7 @@ def seq_to_struct(
         f= max(int((a+b)/6),c+0)
         h= int(1.5*b)
         j = int(1.2*b)
+        k = max(a, len(seq))
 
         reduced_component_distance.append([a,b])
         halfway_component_distance.append([a,d])
@@ -603,8 +608,10 @@ def seq_to_struct(
         g_distance.append([a,h])
         h_distance.append([3,h])
         j_distance.append([a,j])
-        h_distance.append([3,j])
-
+        i_distance.append([3,j])
+        l_distance.append([a,c])
+        k_distance.append([a,b])
+        m_distance.append([a,k])
 
     permutations = get_permutations(position_subsets,component_distance,iii)
     regex_list = []
@@ -620,6 +627,10 @@ def seq_to_struct(
         res6 = build_regex(position_subsets,g_distance, node_dict, convert, positions, mc)
         res7 = build_regex(position_subsets,h_distance, node_dict, convert, positions, mc)
         res8 = build_regex(position_subsets,j_distance, node_dict, convert, positions, mc)
+        res9 = build_regex(position_subsets,i_distance, node_dict, convert, positions, mc)
+        res10 = build_regex(position_subsets,l_distance, node_dict, convert, positions, mc)
+        res11 = build_regex(position_subsets,k_distance, node_dict, convert, positions, mc)
+        res12 = build_regex(position_subsets,m_distance, node_dict, convert, positions, mc)
 
         regex_list.append([res,position_subsets,node_dict])
         regex_list.append([res2,position_subsets,node_dict])
@@ -629,6 +640,10 @@ def seq_to_struct(
         regex_list.append([res6,position_subsets,node_dict])
         regex_list.append([res7,position_subsets,node_dict])
         regex_list.append([res8,position_subsets,node_dict])
+        regex_list.append([res9, position_subsets, node_dict])
+        regex_list.append([res10, position_subsets, node_dict])
+        regex_list.append([res11, position_subsets, node_dict])
+        regex_list.append([res12, position_subsets, node_dict])
     #print("REGEX:",regex_list)
     output = []
     final_model = {}

bayespairing/src/parse_sequences.py

@@ -135,10 +135,11 @@ def run_fasta(input, modules_to_parse, dataset, ss="", arguments={}):
                 seq = str(record.seq)
                 seq = seq.upper()
                 sequences.append(seq)
+                print("LENGTH OF SEQUENCE:",len(seq))
                 #print("PARSING SEQUENCE ", id, "\n")
                 if "T" in seq:
                     seq = str(seq).replace("T", "U")
-                if len(seq) <= 200:
+                if len(seq) <= 300:
                     #print("Running BayesPairing on full sequence")
                     maxs = run_BP(seq, ss, modules_to_parse, dataset, "NONE", m, n, sm, mc, p, k)
                     if interm:
@@ -249,11 +250,12 @@ def run_fasta(input, modules_to_parse, dataset, ss="", arguments={}):
         fOUTPUT= fOUTPUT+"\n"+stats
         pickle.dump(prediction_scores,open("../output/"+o+".pickle","wb"))
     else:
+        print("LENGTH OF SEQUENCE:",len(input))
         if "T" in input:
             input = input.upper()
             input = str(input).replace("T", "U")
         sequences = [input]
-        if len(input) <= 200:
+        if len(input) <= 300:
             maxs = run_BP(input, ss, modules_to_parse, dataset, "NONE", m, n, sm, mc, p, k,sscons)
             all_maxes = []
             for ind, mod in enumerate(modules_to_parse):