Coverage for biobb_vs/utils/box_residues.py: 26%
90 statements
« prev ^ index » next coverage.py v7.15.3, created at 2026-08-03 13:34 +0000
1#!/usr/bin/env python3
3"""Module containing the BoxResidues class and the command line interface."""
4import warnings
5from pathlib import PurePath
6from typing import Optional
7import numpy as np
8from Bio import BiopythonDeprecationWarning
9from biobb_common.generic.biobb_object import BiobbObject
10from biobb_common.tools import file_utils as fu
11from biobb_common.tools.file_utils import launchlogger
13from biobb_vs.utils.common import (
14 _from_string_to_list,
15 check_input_path,
16 check_output_path,
17 get_box_coordinates,
18)
20with warnings.catch_warnings():
21 warnings.simplefilter("ignore", BiopythonDeprecationWarning)
22 # try:
23 # import Bio.SubsMat.MatrixInfo
24 # except ImportError:
25 import Bio.Align.substitution_matrices
26 import Bio.pairwise2
27 import Bio.PDB
30class BoxResidues(BiobbObject):
31 """
32 | biobb_vs BoxResidues
33 | This class sets the center and the size of a rectangular parallelepiped box around a set of residues.
34 | Sets the center and the size of a rectangular parallelepiped box around a selection of residues found in a given PDB. The residue identifiers that compose the selection (i.e. binding site) are provided by a property list.
36 Args:
37 input_pdb_path (str): PDB protein structure for which the box will be build. Its size and center will be set around the 'resid_list' property once mapped against this PDB. File type: input. `Sample file <https://github.com/bioexcel/biobb_vs/raw/master/biobb_vs/test/data/utils/input_box_residues.pdb>`_. Accepted formats: pdb (edam:format_1476).
38 output_pdb_path (str): PDB including the annotation of the box center and size as REMARKs. File type: output. `Sample file <https://github.com/bioexcel/biobb_vs/raw/master/biobb_vs/test/reference/utils/ref_output_box_residues.pdb>`_. Accepted formats: pdb (edam:format_1476).
39 properties (dic - Python dictionary object containing the tool parameters, not input/output files):
40 * **resid_list** (*list*) - (None) List with all the residue numbers to form a cavity or binding site. Mandatory property.
41 * **offset** (*float*) - (2.0) [0.1~1000|0.1] Extra distance (Angstroms) between the last residue atom and the box boundary. The box is centred on the mean of the selected coordinates, so a set of points that is asymmetric about its centre may extend slightly beyond the box faces. Choose an offset large enough to absorb that asymmetry, yet small enough not to enlarge the box past what the binding site needs, since a larger box spreads the same sampling effort over more space.
42 * **box_coordinates** (*bool*) - (False) Add box coordinates as 8 ATOM records.
43 * **residue_offset** (*int*) - (0) [0~1000|1] Residue id offset.
44 * **remove_tmp** (*bool*) - (True) [WF property] Remove temporal files.
45 * **restart** (*bool*) - (False) [WF property] Do not execute if output files exist.
46 * **sandbox_path** (*str*) - ("./") [WF property] Parent path to the sandbox directory.
48 Examples:
49 This is a use example of how to use the building block from Python::
51 from biobb_vs.utils.box_residues import box_residues
52 prop = {
53 'resid_list': [718, 743, 745, 762, 766, 796, 790, 791, 793, 794, 788],
54 'offset': 2,
55 'box_coordinates': True
56 }
57 box_residues(input_pdb_path='/path/to/myStructure.pdb',
58 output_pdb_path='/path/to/newBox.pdb',
59 properties=prop)
61 Info:
62 * wrapped_software:
63 * name: In house using Biopython
64 * version: >=1.76
65 * license: Apache-2.0
66 * ontology:
67 * name: EDAM
68 * schema: http://edamontology.org/EDAM.owl
70 """
72 def __init__(
73 self, input_pdb_path, output_pdb_path, properties=None, **kwargs
74 ) -> None:
75 properties = properties or {}
77 # Call parent class constructor
78 super().__init__(properties)
79 self.locals_var_dict = locals().copy()
81 # Input/Output files
82 self.io_dict = {
83 "in": {"input_pdb_path": input_pdb_path},
84 "out": {"output_pdb_path": output_pdb_path},
85 }
87 # Properties specific for BB
88 self.resid_list = _from_string_to_list(properties.get("resid_list", []))
89 self.offset = float(properties.get("offset", 2.0))
90 self.box_coordinates = float(properties.get("box_coordinates", False))
91 self.residue_offset = properties.get("residue_offset", 0)
92 self.properties = properties
94 # Check the properties
95 self.check_properties(properties)
96 self.check_arguments()
98 def check_data_params(self, out_log, err_log):
99 """Checks all the input/output paths and parameters"""
100 self.io_dict["in"]["input_pdb_path"] = check_input_path(
101 self.io_dict["in"]["input_pdb_path"],
102 "input_pdb_path",
103 self.out_log,
104 self.__class__.__name__,
105 )
106 self.io_dict["out"]["output_pdb_path"] = check_output_path(
107 self.io_dict["out"]["output_pdb_path"],
108 "output_pdb_path",
109 False,
110 self.out_log,
111 self.__class__.__name__,
112 )
114 @launchlogger
115 def launch(self) -> int:
116 """Execute the :class:`BoxResidues <utils.box_residues.BoxResidues>` utils.box_residues.BoxResidues object."""
118 # check input/output paths and parameters
119 self.check_data_params(self.out_log, self.err_log)
121 # Setup Biobb
122 if self.check_restart():
123 return 0
124 self.stage_files()
126 # Parse structure
127 fu.log(
128 "Loading input PDB structure %s" % (self.io_dict["in"]["input_pdb_path"]),
129 self.out_log,
130 self.global_log,
131 )
132 structure_name = PurePath(self.io_dict["in"]["input_pdb_path"]).name
133 parser = Bio.PDB.PDBParser(QUIET=True)
134 structPDB = parser.get_structure(
135 structure_name, self.io_dict["in"]["input_pdb_path"]
136 )
138 if len(structPDB):
139 structPDB = structPDB[0]
141 # Mapping residue structure into input structure
143 fu.log(
144 "Mapping residue structure into input structure",
145 self.out_log,
146 self.global_log,
147 )
149 # Listing residues to be selected from the residue structure
150 residPDB_res_list = []
151 for residPDB_res in self.resid_list:
152 if self.residue_offset:
153 residPDB_res_list.append((" ", residPDB_res + self.residue_offset, " "))
154 else:
155 residPDB_res_list.append((" ", residPDB_res, " "))
157 selection_res_list = []
158 selection_atoms_num = 0
159 for struct_chain in structPDB:
160 for struct_res in struct_chain:
161 if struct_res.get_id() in residPDB_res_list:
162 selection_res_list.append(struct_res)
163 selection_atoms_num += len(struct_res.get_list())
165 if len(selection_res_list) == 0:
166 fu.log(
167 self.__class__.__name__ + ": Cannot match any of the residues listed in [%s] into %s"
168 % (
169 ", ".join(str(v) for v in self.resid_list),
170 self.io_dict["in"]["input_pdb_path"],
171 ),
172 self.out_log,
173 )
174 raise SystemExit(
175 self.__class__.__name__ + ": Cannot match any of the residues listed in [%s] into %s"
176 % (
177 ", ".join(str(v) for v in self.resid_list),
178 self.io_dict["in"]["input_pdb_path"],
179 )
180 )
181 elif len(selection_res_list) != len(residPDB_res_list):
182 fu.log(
183 "Cannot match all the residues listed in %s into %s. Found %s out of %s"
184 % (
185 ", ".join(str(v) for v in self.resid_list),
186 self.io_dict["in"]["input_pdb_path"],
187 len(selection_res_list),
188 len(residPDB_res_list),
189 ),
190 self.out_log,
191 )
192 else:
193 fu.log(
194 "Selection of residues successfully matched",
195 self.out_log,
196 self.global_log,
197 )
199 # Compute binding site box size
201 # compute box center
202 selection_box_center = (
203 sum(atom.coord for res in selection_res_list for atom in res.get_atoms()) / selection_atoms_num
204 )
205 fu.log(
206 "Binding site center (Angstroms): %10.3f%10.3f%10.3f"
207 % (
208 selection_box_center[0],
209 selection_box_center[1],
210 selection_box_center[2],
211 ),
212 self.out_log,
213 self.global_log,
214 )
216 # compute box size
217 selection_coords_max = np.amax(
218 [atom.coord for res in selection_res_list for atom in res.get_atoms()],
219 axis=0,
220 )
221 selection_box_size = selection_coords_max - selection_box_center
222 if self.offset:
223 selection_box_size = [c + self.offset for c in selection_box_size]
225 # SIZE is written as the full edge length, which is what Vina reads from
226 # --size_x/y/z. selection_box_size stays a half-extent because that is
227 # what the corner coordinates and the volume are computed from.
228 selection_box_edge = [2 * side for side in selection_box_size]
229 fu.log(
230 "Binding site box edge lengths (Angstroms): %10.3f%10.3f%10.3f"
231 % (selection_box_edge[0], selection_box_edge[1], selection_box_edge[2]),
232 self.out_log,
233 self.global_log,
234 )
236 # compute volume
237 vol = np.prod(selection_box_size) * 2**3
238 fu.log("Volume (cubic Angstroms): %.0f" % (vol), self.out_log, self.global_log)
240 # add box details as PDB remarks
241 remarks = "REMARK BOX CENTER:%10.3f%10.3f%10.3f" % (
242 selection_box_center[0],
243 selection_box_center[1],
244 selection_box_center[2],
245 )
246 remarks += " SIZE:%10.3f%10.3f%10.3f" % (
247 selection_box_edge[0],
248 selection_box_edge[1],
249 selection_box_edge[2],
250 )
252 selection_box_coords_txt = ""
253 # add (optional) box coordinates as 8 ATOM records
254 if self.box_coordinates:
255 fu.log("Adding box coordinates", self.out_log, self.global_log)
256 selection_box_coords_txt = get_box_coordinates(
257 selection_box_center, selection_box_size
258 )
260 with open(self.io_dict["out"]["output_pdb_path"], "w") as f:
261 f.seek(0, 0)
262 f.write(remarks.rstrip("\r\n") + "\n" + selection_box_coords_txt)
264 fu.log(
265 "Saving output PDB file (with box setting annotations): %s"
266 % (self.io_dict["out"]["output_pdb_path"]),
267 self.out_log,
268 self.global_log,
269 )
271 # Copy files to host
272 self.copy_to_host()
273 self.remove_tmp_files()
275 return 0
278def box_residues(
279 input_pdb_path: str,
280 output_pdb_path: str,
281 properties: Optional[dict] = None,
282 **kwargs,
283) -> int:
284 """Create the :class:`BoxResidues <utils.box_residues.BoxResidues>` class and
285 execute the :meth:`launch() <utils.box_residues.BoxResidues.launch>` method."""
286 return BoxResidues(**dict(locals())).launch()
289box_residues.__doc__ = BoxResidues.__doc__
290main = BoxResidues.get_main(box_residues, "Sets the center and the size of a rectangular parallelepiped box around a selection of residues found in a given PDB.")
293if __name__ == "__main__":
294 main()