Create Slice Syntax for tied_list for make_tied_positions_dict.py

examples/submit_example_5_new_protocol.sh

@@ -0,0 +1,41 @@
+#!/bin/bash
+#SBATCH -p gpu
+#SBATCH --mem=32g
+#SBATCH --gres=gpu:rtx2080:1
+#SBATCH -c 3
+#SBATCH --output=example_5.out
+
+source activate mlfold
+
+folder_with_pdbs="../inputs/PDB_test_enzymes/8E0X/"
+
+output_dir="../outputs/example_5_new_protocol_outputs"
+if [ ! -d $output_dir ]
+then
+    mkdir -p $output_dir
+fi
+
+
+path_for_parsed_chains=$output_dir"/parsed_pdbs.jsonl"
+path_for_assigned_chains=$output_dir"/assigned_pdbs.jsonl"
+path_for_fixed_positions=$output_dir"/fixed_pdbs.jsonl"
+path_for_tied_positions=$output_dir"/tied_pdbs.jsonl"
+chains_to_design="A B C D"
+tied_positions="5:310 315:347, 2:307 312:344, 2:307 312:344, 1:306 311:343" #two list must match in length; residue 1 in chain A and C will be sampled togther;
+
+python ../helper_scripts/parse_multiple_chains.py --input_path=$folder_with_pdbs --output_path=$path_for_parsed_chains
+
+python ../helper_scripts/assign_fixed_chains.py --input_path=$path_for_parsed_chains --output_path=$path_for_assigned_chains --chain_list "$chains_to_design"
+
+python ../helper_scripts/make_tied_positions_dict.py --input_path=$path_for_parsed_chains --output_path=$path_for_tied_positions --chain_list "$chains_to_design" --position_list "$tied_positions"
+
+python ../protein_mpnn_run.py \
+        --jsonl_path $path_for_parsed_chains \
+        --chain_id_jsonl $path_for_assigned_chains \
+        --fixed_positions_jsonl $path_for_fixed_positions \
+        --tied_positions_jsonl $path_for_tied_positions \
+        --out_folder $output_dir \
+        --num_seq_per_target 2 \
+        --sampling_temp "0.1" \
+        --seed 37 \
+        --batch_size 1

helper_scripts/make_tied_positions_dict.py

@@ -1,5 +1,19 @@
 import argparse
 
+def parse_tied_str(tied_position_str):
+    tied_list = []
+    for one in tied_position_str.split(","):
+        one_list = []
+        for item in one.split():
+            if len(item.split(":")) == 1:
+                one_list.append(int(item.split(":")))
+            elif len(item.split(":")) == 2:
+                one_list.extend(list(range(int(item.split(":")[0]), int(item.split(":")[1]))))
+            else:
+                raise RuntimeError("Now we only support slice syntax(1:10, 2:11) and ordinary syntax(1 2, 2 3)")
+        tied_list.append(one_list)
+    return tied_list
+
 def main(args):
 
     import glob
@@ -14,7 +28,7 @@ def main(args):
     homooligomeric_state = args.homooligomer
 
     if homooligomeric_state == 0:
-        tied_list = [[int(item) for item in one.split()] for one in args.position_list.split(",")]
+        tied_list = parse_tied_str(args.position_list)
         global_designed_chain_list = [str(item) for item in args.chain_list.split()]
         my_dict = {}
         for json_str in json_list:
@@ -49,7 +63,7 @@ def main(args):
     argparser.add_argument("--input_path", type=str, help="Path to the parsed PDBs")
     argparser.add_argument("--output_path", type=str, help="Path to the output dictionary")
     argparser.add_argument("--chain_list", type=str, default='', help="List of the chains that need to be fixed")
-    argparser.add_argument("--position_list", type=str, default='', help="Position lists, e.g. 11 12 14 18, 1 2 3 4 for first chain and the second chain")
+    argparser.add_argument("--position_list", type=str, default='', help="Position lists, e.g. 11 12 14 18, 1 2 3 4 for first chain and the second chain. Slice is now supported.")
     argparser.add_argument("--homooligomer", type=int, default=0, help="If 0 do not use, if 1 then design homooligomer")
 
     args = argparser.parse_args()