Formatting

from rdkit import Chem

import deepchem as dc

def load_pdbbind_labels(labels_file):

  """Loads pdbbind labels as dataframe"""

  # Some complexes have labels but no PDB files. Filter these manually

  missing_pdbs = ["1d2v", "1jou", "1s8j", "3f39", "3i3d", "3i3b", "3dyo", "3t0d", "1cam", "3vdb", "3f37", "3f38", "4mlt", "3f36", "4o7d", "3t08", "3f34", "3f35", "2wik", "4mlx", "2wij", "1px4", "4wkt", "3f33", "2wig", "3muz", "3t2p", "3t2q", "4pji", "2adj", "3t09", "3mv0", "1pts", "3vd9", "3axk", "4q1s", "3t0b", "4b82", "3vd7", "3hg1", "3vd4", "3vdc", "3b5y", "4oi6", "3axm", "4mdm", "2mlm", "3eql", "4ob0", "3wi6", "4fgt", "4pnc", "4mvn", "4lv3", "4lz9", "1pyg", "3h1k", "7gpb", "1e8h", "4wku", "2f2h", "1zyr", "1z9j", "3b5d", "3b62", "4q3q", "4mdl", "4no6", "4mdg", "3dxj", "4u0x", "4l6q", "4q3r", "1h9s", "4ob1", "4ob2", "4qq5", "4nk3", "3k1j", "4m8t", "4mzo", "4nnn", "4q3s", "4nnw", "3cf1", "4u5t", "4wkv", "4ool", "3a2c", "4wm9", "4pkb", "4qkx", "4no8", "1ztz", "1nu1", "4kn4", "4mao", "4qqc", "4len", "4lv1", "4r02", "4r6v", "4fil", "4q2k", "1hpb", "4oon", "4qbb", "4ruu", "4no1", "3w8o", "4kn7", "4r17", "4r18", "5hvp", "1e59", "1sqq", "3n75", "4kmu", "4mzs", "1sqb", "1lr8", "4lv2", "4wmc", "1sqp", "3whw", "4cpa", "3i8w", "4hrd", "4hrc", "1ntk", "1rbo"]

  missing_pdbs = [

      "1d2v", "1jou", "1s8j", "3f39", "3i3d", "3i3b", "3dyo", "3t0d", "1cam",

      "3vdb", "3f37", "3f38", "4mlt", "3f36", "4o7d", "3t08", "3f34", "3f35",

      "2wik", "4mlx", "2wij", "1px4", "4wkt", "3f33", "2wig", "3muz", "3t2p",

      "3t2q", "4pji", "2adj", "3t09", "3mv0", "1pts", "3vd9", "3axk", "4q1s",

      "3t0b", "4b82", "3vd7", "3hg1", "3vd4", "3vdc", "3b5y", "4oi6", "3axm",

      "4mdm", "2mlm", "3eql", "4ob0", "3wi6", "4fgt", "4pnc", "4mvn", "4lv3",

      "4lz9", "1pyg", "3h1k", "7gpb", "1e8h", "4wku", "2f2h", "1zyr", "1z9j",

      "3b5d", "3b62", "4q3q", "4mdl", "4no6", "4mdg", "3dxj", "4u0x", "4l6q",

      "4q3r", "1h9s", "4ob1", "4ob2", "4qq5", "4nk3", "3k1j", "4m8t", "4mzo",

      "4nnn", "4q3s", "4nnw", "3cf1", "4u5t", "4wkv", "4ool", "3a2c", "4wm9",

      "4pkb", "4qkx", "4no8", "1ztz", "1nu1", "4kn4", "4mao", "4qqc", "4len",

      "4lv1", "4r02", "4r6v", "4fil", "4q2k", "1hpb", "4oon", "4qbb", "4ruu",

      "4no1", "3w8o", "4kn7", "4r17", "4r18", "5hvp", "1e59", "1sqq", "3n75",

      "4kmu", "4mzs", "1sqb", "1lr8", "4lv2", "4wmc", "1sqp", "3whw", "4cpa",

      "3i8w", "4hrd", "4hrc", "1ntk", "1rbo"

  ]

  contents = []

  with open(labels_file) as f:

    for line in f:

               "ignore-this-field", "reference", "ligand name"))

  return contents_df

def compute_pdbbind_features(grid_featurizer, pdb_subdir, pdb_code):

  """Compute features for a given complex"""

  protein_file = os.path.join(pdb_subdir, "%s_protein.pdb" % pdb_code)

  ligand_file = os.path.join(pdb_subdir, "%s_ligand.sdf" % pdb_code)

  features = grid_featurizer.featurize_complexes(

    [ligand_file], [protein_file])

  features = grid_featurizer.featurize_complexes([ligand_file], [protein_file])

  features = np.squeeze(features)

  return features

def featurize_pdbbind(data_dir=None, feat="grid", subset="core"):

  """Featurizes pdbbind according to provided featurization"""

  tasks = ["-logKd/Ki"]

  # Define featurizers

  if feat == "grid":

    featurizer = dc.feat.GridFeaturizer(

        voxel_width=16.0, feature_types="voxel_combined",

        voxel_width=16.0,

        feature_types="voxel_combined",

        # TODO(rbharath, enf, leswing): Figure out why pi_stack and cation_pi

        # reduce validation performance

        # voxel_feature_types=["ecfp", "splif", "hbond", "pi_stack", "cation_pi",

        # "salt_bridge"], ecfp_power=9, splif_power=9,

        voxel_feature_types=["ecfp", "splif", "hbond", "salt_bridge"],

        ecfp_power=9, splif_power=9,

        parallel=True, flatten=True)

        ecfp_power=9,

        splif_power=9,

        parallel=True,

        flatten=True)

  elif feat == "coord":

    neighbor_cutoff = 4

    max_num_neighbors = 10

    featurizer = dc.feat.NeighborListComplexAtomicCoordinates(

        max_num_neighbors, neighbor_cutoff)

    featurizer = dc.feat.NeighborListComplexAtomicCoordinates(max_num_neighbors,

                                                              neighbor_cutoff)

  else:

    raise ValueError("feat not defined.")

      print("%s is missing!" % pdb_subdir)

      missing_pdbs.append(pdb_subdir)

      continue

    computed_feature = compute_pdbbind_features(

        featurizer, pdb_subdir, pdb_code)

    computed_feature = compute_pdbbind_features(featurizer, pdb_subdir,

                                                pdb_code)

    if feature_len is None:

      feature_len = len(computed_feature)

    if len(computed_feature) != feature_len:

  dataset = dc.data.DiskDataset.from_numpy(X, y, w, ids, data_dir=data_dir)

  return dataset, tasks

def load_pdbbind_grid(split="index", featurizer="grid", subset="full"):

  """Load PDBBind datasets. Does not do train/test split"""

  dataset, tasks = featurize_pdbbind(feat=featurizer, subset=subset)

  splitters = {'index': dc.splits.IndexSplitter(),

               'random': dc.splits.RandomSplitter()}

  splitters = {

      'index': dc.splits.IndexSplitter(),

      'random': dc.splits.RandomSplitter()

  }

  splitter = splitters[split]

  train, valid, test = splitter.train_valid_test_split(dataset)