File size: 20,543 Bytes
6d19a18 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 | #!/usr/bin/env python3
"""Analyse docking poses with PoseBusters and OST.
The script processes one docking run directory (for example
``docking/gnina_redock``) and writes pose-level CSV and Parquet files beneath a
shared analysis directory.
"""
from __future__ import annotations
import argparse
import csv
import json
import logging
import re
import shutil
import subprocess
from concurrent.futures import ProcessPoolExecutor, as_completed
from pathlib import Path
from typing import Sequence
import pandas as pd
from posebusters import PoseBusters
from rdkit import Chem, RDLogger
SUPPORTED_METHODS = {"gnina", "smina", "diffdock"}
def file_ok(path: Path) -> bool:
"""Return whether a file exists and is non-empty."""
return path.is_file() and path.stat().st_size > 0
def infer_similarity_label(path: Path) -> str:
"""Infer a date label from a similarity TSV filename."""
match = re.search(r"\d{4}-\d{2}-\d{2}", path.name)
if match is None:
raise ValueError(f"Could not infer date label from similarity TSV: {path}")
return match.group(0)
def load_similarity_dicts(
similarity_tsvs: Sequence[Path] | None,
) -> dict[str, dict[str, str]] | None:
"""Load optional similarity TSVs and suffix metric names by date."""
if not similarity_tsvs:
return None
metric_cols = ["sucos_shape", "pocket_qcov", "sucos_shape_pocket_qcov"]
similarity_dict: dict[str, dict[str, str]] = {}
for similarity_tsv in similarity_tsvs:
path = similarity_tsv.expanduser().resolve()
if not file_ok(path):
raise FileNotFoundError(f"Similarity TSV not found or empty: {path}")
label = infer_similarity_label(path)
with path.open("r", encoding="utf-8", newline="") as handle:
reader = csv.DictReader(handle, delimiter="\t")
if reader.fieldnames is None or reader.fieldnames[0] is None:
raise ValueError(f"Similarity TSV has no header: {path}")
missing = set(metric_cols) - set(reader.fieldnames)
if missing:
raise ValueError(
f"Similarity TSV is missing columns {sorted(missing)}: {path}"
)
id_column = reader.fieldnames[0]
for row in reader:
complex_id = row[id_column]
similarity_dict.setdefault(complex_id, {})
for col in metric_cols:
similarity_dict[complex_id][f"{col}-{label}"] = row[col]
return similarity_dict
def add_similarity_metrics(
df: pd.DataFrame,
similarity_dict: dict[str, dict[str, str]] | None,
) -> pd.DataFrame:
"""Attach optional similarity metrics for each comparison complex."""
if similarity_dict is None:
return df
metric_cols = sorted({col for values in similarity_dict.values() for col in values})
for col in metric_cols:
df[col] = df["compare_id"].map(
lambda complex_id: similarity_dict.get(complex_id, {}).get(col)
)
return df
def structure_root_from_run(run_dir: Path) -> Path:
"""Resolve the dataset root from ``<dataset>/docking/<run>``."""
if run_dir.parent.name != "docking":
raise ValueError(
f"Run directory must be directly beneath a 'docking' directory: {run_dir}"
)
structure_root = run_dir.parent.parent
if not (structure_root / "prepared_inputs" / "redock").is_dir():
raise FileNotFoundError(
f"Redock prepared inputs not found beneath dataset root: {structure_root}"
)
return structure_root
def reference_paths(structure_root: Path, complex_id: str) -> tuple[Path, Path]:
"""Return the redock receptor and crystallographic ligand for one complex."""
directory = structure_root / "prepared_inputs" / "redock" / complex_id
protein = directory / f"{complex_id}_prepared.pdb"
ligand = directory / f"{complex_id}_reference_ligand.sdf"
if not file_ok(protein):
raise FileNotFoundError(f"Missing reference receptor: {protein}")
if not file_ok(ligand):
raise FileNotFoundError(f"Missing reference ligand: {ligand}")
return protein, ligand
def docked_protein_path(structure_root: Path, mode: str, complex_id: str) -> Path:
"""Return the exact receptor used for one docking mode."""
directory = structure_root / "prepared_inputs" / mode / complex_id
if not directory.is_dir():
raise FileNotFoundError(f"Prepared receptor directory not found: {directory}")
if mode == "redock":
path = directory / f"{complex_id}_prepared.pdb"
elif mode == "fragment_crossdock":
matches = sorted(directory.glob("fragment_crossdock_*_aligned.pdb"))
if len(matches) != 1:
raise FileNotFoundError(
f"Expected one fragment-crossdock receptor in {directory}; found {len(matches)}"
)
path = matches[0]
else:
path = directory / f"{mode}_prepared_aligned.pdb"
if not file_ok(path):
raise FileNotFoundError(f"Docked receptor not found or empty: {path}")
return path
def seed_dirs(complex_dir: Path) -> list[Path]:
"""Return immediate seed directories for one docked complex."""
directories = sorted(path for path in complex_dir.glob("seed_*") if path.is_dir())
if not directories:
raise FileNotFoundError(f"No seed directories found in: {complex_dir}")
return directories
def pose_dir(seed_dir: Path) -> Path:
"""Return the canonical docked-pose directory."""
directory = seed_dir / "docked_poses"
if not directory.is_dir():
raise FileNotFoundError(f"Docked-pose directory not found: {directory}")
return directory
def format_chain_mapping(mapping: dict | None) -> str | None:
"""Format an OST chain mapping as a compact string."""
if not mapping:
return None
return ",".join(f"{key}:{value}" for key, value in sorted(mapping.items()))
def ost_to_df(path: Path) -> pd.DataFrame:
"""Convert OST ligand-comparison output to a pose-level table."""
with path.open("r", encoding="utf-8") as handle:
ost = json.load(handle)
lddt_rows = []
for row in ost["lddt_pli"]["full_results"]:
sample = int(row["model_ligand"].rsplit(":", 1)[1])
score = row.get("score")
lddt_rows.append(
{"sample": sample, "lddt_pli": float(score) if score is not None else 0.0}
)
rmsd_rows = []
for row in ost["rmsd"]["full_results"]:
sample = int(row["model_ligand"].rsplit(":", 1)[1])
score = row.get("score")
rmsd_rows.append(
{
"sample": sample,
"lig_rmsd": float(score) if score is not None else 99.0,
"rmsd_chain_mapping": format_chain_mapping(row.get("chain_mapping")),
"lddt_lp": row.get("lddt_lp"),
"pocket_bb_rmsd": row.get("bb_rmsd"),
}
)
return pd.DataFrame(lddt_rows).merge(
pd.DataFrame(rmsd_rows), on="sample", how="outer"
)
def posebusters_to_df(path: Path) -> pd.DataFrame:
"""Convert PoseBusters output to a pose-level validity table."""
df = pd.read_csv(path).copy()
if "position" in df.columns:
df["sample"] = df["position"].astype(int)
else:
df["sample"] = range(len(df))
meta_cols = {"file", "molecule", "position", "sample"}
metric_cols = [col for col in df.columns if col not in meta_cols]
if not metric_cols:
raise ValueError(f"PoseBusters output contains no check columns: {path}")
df["pb_valid"] = df[metric_cols].all(axis=1)
return df[["sample", "pb_valid"]]
def sdf_to_pose_df(
path: Path, score_property: str, flip_sign: bool = False
) -> pd.DataFrame:
"""Read pose scores from a ranked SDF file."""
rows: list[dict[str, object]] = []
for sample, mol in enumerate(Chem.SDMolSupplier(str(path), removeHs=False)):
if mol is None:
raise ValueError(f"Invalid molecule at position {sample} in {path}")
if not mol.HasProp(score_property):
raise ValueError(f"Pose {sample} in {path} is missing '{score_property}'")
score = float(mol.GetProp(score_property))
rows.append({"sample": sample, "rank_score": -score if flip_sign else score})
if not rows:
raise ValueError(f"No poses found in SDF: {path}")
return pd.DataFrame(rows)
def rank_molecules(
molecules: list[Chem.Mol],
score_property: str,
descending: bool,
) -> list[Chem.Mol]:
"""Rank molecules by an SDF score property and annotate their rank."""
for molecule in molecules:
if not molecule.HasProp(score_property):
raise ValueError(
f"Molecule is missing required score property '{score_property}'"
)
ranked = sorted(
molecules,
key=lambda molecule: float(molecule.GetProp(score_property)),
reverse=descending,
)
for rank, molecule in enumerate(ranked, start=1):
molecule.SetIntProp("rank", rank)
return ranked
def join_diffdock_sdfs(input_dir: Path, output_sdf: Path) -> None:
"""Combine DiffDock rank files while preserving confidence and rank metadata."""
files = sorted(input_dir.glob("rank*_confidence*.sdf"))
if not files:
raise FileNotFoundError(f"No DiffDock SDF files found in: {input_dir}")
def parse_rank_conf(path: Path) -> tuple[int, float]:
match = re.search(r"rank(\d+)_confidence([-\d\.eE]+)", path.stem)
if match is None:
raise ValueError(f"Unexpected DiffDock filename: {path.name}")
return int(match.group(1)), float(match.group(2))
with Chem.SDWriter(str(output_sdf)) as writer:
for path in sorted(files, key=lambda item: parse_rank_conf(item)[0]):
molecules = [mol for mol in Chem.SDMolSupplier(str(path), removeHs=False)]
if len(molecules) != 1 or molecules[0] is None:
raise ValueError(f"Expected exactly one valid molecule in {path}")
rank, confidence = parse_rank_conf(path)
molecule = molecules[0]
molecule.SetIntProp("rank", rank)
molecule.SetProp("confidence", f"{confidence:.6f}")
writer.write(molecule)
def join_docking_sdfs(
input_dir: Path,
output_sdf: Path,
score_property: str,
descending: bool,
) -> None:
"""Combine all GNINA/Smina pose files and rank poses globally."""
files = sorted(input_dir.glob("*_docked_*.sdf"))
if not files:
raise FileNotFoundError(f"No docking SDF files found in: {input_dir}")
molecules: list[Chem.Mol] = []
for path in files:
for index, molecule in enumerate(Chem.SDMolSupplier(str(path), removeHs=False)):
if molecule is None:
raise ValueError(f"Invalid molecule at position {index} in {path}")
molecules.append(molecule)
with Chem.SDWriter(str(output_sdf)) as writer:
for molecule in rank_molecules(molecules, score_property, descending):
writer.write(molecule)
def run_posebusters(
buster: PoseBusters,
sdf_path: Path,
receptor_path: Path,
output_path: Path,
overwrite: bool,
) -> None:
"""Run PoseBusters unless a valid cached output is available."""
if file_ok(output_path) and not overwrite:
return
molecules = list(Chem.SDMolSupplier(str(sdf_path), removeHs=False))
if not molecules or any(molecule is None for molecule in molecules):
raise ValueError(f"Could not read all poses from {sdf_path}")
results = buster.bust(molecules, None, receptor_path)
results.to_csv(output_path, index=False)
def run_ost(
sdf_path: Path,
docked_receptor_path: Path,
reference_receptor_path: Path,
reference_ligand_path: Path,
output_path: Path,
overwrite: bool,
) -> None:
"""Run OST ligand-structure comparison unless cached output is available."""
if file_ok(output_path) and not overwrite:
return
command = [
"ost",
"compare-ligand-structures",
"-m",
str(docked_receptor_path),
"-ml",
str(sdf_path),
"-r",
str(reference_receptor_path),
"-rl",
str(reference_ligand_path),
"--lddt-pli",
"--rmsd",
"-o",
str(output_path),
"-v",
"0",
"--full-results",
]
subprocess.run(command, check=True)
def process_complex(
complex_dir: Path,
structure_root: Path,
method: str,
mode: str,
run_label: str,
overwrite: bool,
similarity_dict: dict[str, dict[str, str]] | None,
) -> pd.DataFrame:
"""Analyse all seeds for one complex."""
complex_id = complex_dir.name
RDLogger.DisableLog("rdApp.warning")
buster = PoseBusters(config="dock")
reference_receptor, reference_ligand = reference_paths(structure_root, complex_id)
docked_receptor = docked_protein_path(structure_root, mode, complex_id)
reference_mol = Chem.MolFromMolFile(str(reference_ligand), removeHs=False)
if reference_mol is None:
raise ValueError(f"Could not read reference ligand: {reference_ligand}")
ligand_smiles = Chem.MolToSmiles(reference_mol)
if method == "gnina":
score_property, flip_sign, descending = "CNNscore", False, True
elif method == "smina":
score_property, flip_sign, descending = "minimizedAffinity", True, False
elif method == "diffdock":
score_property, flip_sign, descending = "confidence", False, True
else:
raise ValueError(f"Unsupported method: {method}")
results: list[pd.DataFrame] = []
for seed_dir in seed_dirs(complex_dir):
raw_pose_dir = pose_dir(seed_dir)
joined_sdf = seed_dir / f"{complex_id}_{run_label}_joined.sdf"
if method == "diffdock":
if not file_ok(joined_sdf) or overwrite:
join_diffdock_sdfs(raw_pose_dir, joined_sdf)
else:
if not file_ok(joined_sdf) or overwrite:
join_docking_sdfs(raw_pose_dir, joined_sdf, score_property, descending)
pb_path = seed_dir / f"PB_{complex_id}_{run_label}.csv"
run_posebusters(buster, joined_sdf, docked_receptor, pb_path, overwrite)
ost_path = seed_dir / f"OST_{complex_id}_{run_label}.json"
run_ost(
joined_sdf,
docked_receptor,
reference_receptor,
reference_ligand,
ost_path,
overwrite,
)
pose_df = sdf_to_pose_df(joined_sdf, score_property, flip_sign=flip_sign)
pb_df = posebusters_to_df(pb_path)
ost_df = ost_to_df(ost_path)
ost_df["compare_id"] = complex_id
df = pose_df.merge(pb_df, on="sample", how="left", validate="one_to_one").merge(
ost_df, on="sample", how="left", validate="one_to_one"
)
df = add_similarity_metrics(df, similarity_dict)
df["dock_prot"] = mode
df["complex_id"] = complex_id
df["seed"] = seed_dir.name
df["method"] = method
df["ligand_smiles"] = ligand_smiles
results.append(df)
return pd.concat(results, ignore_index=True)
def build_parser() -> argparse.ArgumentParser:
"""Build the command-line parser."""
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument(
"run_dir", type=Path, help="One method/mode docking run directory."
)
parser.add_argument("--method", required=True, choices=sorted(SUPPORTED_METHODS))
parser.add_argument("--mode", required=True, help="Prepared docking mode name.")
parser.add_argument(
"--similarity-tsv",
type=Path,
nargs="+",
default=None,
help="Optional similarity TSV file(s).",
)
parser.add_argument(
"--num-workers",
"-j",
type=int,
default=8,
help="Number of complexes analysed in parallel (default: 8).",
)
parser.add_argument(
"--overwrite",
action="store_true",
help="Recompute cached OST/PoseBusters outputs.",
)
parser.add_argument(
"--output-dir",
type=Path,
default=None,
help="Analysis root (default: sibling analysis directory).",
)
parser.add_argument("--suffix", default="", help="Optional output filename suffix.")
parser.add_argument("--verbose", "-v", action="store_true")
return parser
def main() -> None:
"""Run docking analysis."""
args = build_parser().parse_args()
logging.basicConfig(
level=logging.DEBUG if args.verbose else logging.INFO,
format="%(asctime)s %(levelname)s %(message)s",
datefmt="%Y-%m-%d %H:%M:%S",
)
logging.getLogger("posebusters").setLevel(logging.ERROR)
RDLogger.DisableLog("rdApp.warning")
if shutil.which("ost") is None:
raise FileNotFoundError("OST executable 'ost' was not found on PATH")
if args.num_workers < 1:
raise ValueError("--num-workers must be at least 1")
run_dir = args.run_dir.expanduser().resolve()
if not run_dir.is_dir():
raise FileNotFoundError(f"Docking run directory not found: {run_dir}")
expected_name = f"{args.method}_{args.mode}"
if run_dir.name != expected_name:
raise ValueError(
f"Run directory name must be '{expected_name}', got '{run_dir.name}'"
)
structure_root = structure_root_from_run(run_dir)
output_dir = (
args.output_dir.expanduser().resolve()
if args.output_dir is not None
else run_dir.parent.parent / "analysis"
)
similarity_dict = load_similarity_dicts(args.similarity_tsv)
run_label = run_dir.name
complex_dirs = sorted(path for path in run_dir.iterdir() if path.is_dir())
if not complex_dirs:
raise ValueError(f"No complex directories found in: {run_dir}")
all_results: list[pd.DataFrame] = []
failures: list[tuple[str, str]] = []
jobs = [
(
complex_dir,
structure_root,
args.method,
args.mode,
run_label,
args.overwrite,
similarity_dict,
)
for complex_dir in complex_dirs
]
with ProcessPoolExecutor(max_workers=args.num_workers) as executor:
futures = {executor.submit(process_complex, *job): job[0].name for job in jobs}
for future in as_completed(futures):
complex_id = futures[future]
try:
all_results.append(future.result())
logging.info("Finished %s", complex_id)
except Exception as exc:
failures.append((complex_id, str(exc)))
logging.exception("Failed %s", complex_id)
if not all_results:
raise RuntimeError("Analysis produced no results")
final_df = pd.concat(all_results, ignore_index=True)
column_order = [
"complex_id",
"dock_prot",
"method",
"seed",
"sample",
"rank_score",
"pb_valid",
"lig_rmsd",
"lddt_pli",
"compare_id",
"lddt_lp",
"pocket_bb_rmsd",
"rmsd_chain_mapping",
"ligand_smiles",
]
existing = [column for column in column_order if column in final_df.columns]
extra = [column for column in final_df.columns if column not in column_order]
final_df = final_df[existing + extra]
suffix = args.suffix
if suffix and not suffix.startswith("_"):
suffix = f"_{suffix}"
base_name = f"{run_label}{suffix}"
csv_dir = output_dir / "csv"
parquet_dir = output_dir / "parquet"
csv_dir.mkdir(parents=True, exist_ok=True)
parquet_dir.mkdir(parents=True, exist_ok=True)
final_df.to_csv(csv_dir / f"{base_name}.csv", index=False)
final_df.to_parquet(parquet_dir / f"{base_name}.parquet", index=False)
logging.info("Wrote %d pose rows for %s", len(final_df), run_label)
if failures:
failure_path = output_dir / "logs" / f"{run_label}_failed_complexes.tsv"
failure_path.parent.mkdir(parents=True, exist_ok=True)
with failure_path.open("w", encoding="utf-8") as handle:
handle.write("complex_id\treason\n")
for complex_id, reason in failures:
handle.write(
f"{complex_id}\t{reason.replace(chr(9), ' ').replace(chr(10), ' ')}\n"
)
raise SystemExit(f"Analysis completed with {len(failures)} failed complexes")
if __name__ == "__main__":
main()
|