Loading deepchem/feat/contact_fingerprints.py +17 −6 Original line number Diff line number Diff line Loading @@ -133,8 +133,10 @@ class ContactCircularVoxelizer(ComplexFeaturizer): in which they originated. Featurizes a macromolecular complex into a tensor of shape `(voxels_per_edge, voxels_per_edge, voxels_per_edge, size)` where `voxels_per_edge = int(box_width/voxel_width)`. `(voxels_per_edge, voxels_per_edge, voxels_per_edge, size)` where `voxels_per_edge = int(box_width/voxel_width)`. If `flatten==True`, then returns a flattened version of this tensor of length `size*voxels_per_edge**3` """ def __init__(self, Loading @@ -142,7 +144,8 @@ class ContactCircularVoxelizer(ComplexFeaturizer): radius=2, size=8, box_width=16.0, voxel_width=1.0): voxel_width=1.0, flatten=False): """ Parameters ---------- Loading @@ -157,6 +160,9 @@ class ContactCircularVoxelizer(ComplexFeaturizer): is centered on a ligand centroid. voxel_width: float, optional (default 1.0) Size of a 3D voxel in a grid. flatten: bool, optional (default False) If True, then returns a flat feature vector rather than voxel grid. This feature vector is constructed by flattening the usual voxel grid. """ self.cutoff = cutoff self.radius = radius Loading @@ -164,6 +170,7 @@ class ContactCircularVoxelizer(ComplexFeaturizer): self.box_width = box_width self.voxel_width = voxel_width self.voxels_per_edge = int(self.box_width / self.voxel_width) self.flatten = flatten def _featurize_complex(self, molecular_complex): """ Loading Loading @@ -193,7 +200,6 @@ class ContactCircularVoxelizer(ComplexFeaturizer): sum([ voxelize( convert_atom_to_voxel, self.voxels_per_edge, self.box_width, self.voxel_width, hash_ecfp, Loading @@ -209,5 +215,10 @@ class ContactCircularVoxelizer(ComplexFeaturizer): ecfp_degree=self.radius)) ]) ) # Features are of shape (voxels_per_edge, voxels_per_edge, voxels_per_edge, num_feat) so we should concatenate on the last axis. if self.flatten: return np.concatenate([features.flatten() for features in pairwise_features]) else: # Features are of shape (voxels_per_edge, voxels_per_edge, # voxels_per_edge, num_feat) so we should concatenate on the last # axis. return np.concatenate(pairwise_features, axis=-1) deepchem/feat/rdkit_grid_featurizer.py +6 −8 Original line number Diff line number Diff line Loading @@ -37,8 +37,6 @@ logger = logging.getLogger(__name__) http://stackoverflow.com/questions/2827393/angles-between-two-n-dimensional-vectors-in-python """ # TODO(rbharath): Consider this comment on rdkit forums https://github.com/rdkit/rdkit/issues/1590 about sybyl featurization FLAT_FEATURES = ['ecfp_ligand', 'ecfp_hashed', 'splif_hashed', 'hbond_count'] VOXEL_FEATURES = [ Loading Loading @@ -306,12 +304,12 @@ class RdkitGridFeaturizer(ComplexFeaturizer): logger.warning("Some molecules cannot be loaded by Rdkit. Skipping") return None time1 = time.time() centroid = compute_centroid(ligand_xyz) ligand_xyz = subtract_centroid(ligand_xyz, centroid) protein_xyz = subtract_centroid(protein_xyz, centroid) time2 = time.time() logger.info("TIMING: Centroid processing took %0.3f s" % (time2 - time1)) #time1 = time.time() #centroid = compute_centroid(ligand_xyz) #ligand_xyz = subtract_centroid(ligand_xyz, centroid) #protein_xyz = subtract_centroid(protein_xyz, centroid) #time2 = time.time() #logger.info("TIMING: Centroid processing took %0.3f s" % (time2 - time1)) pairwise_distances = compute_pairwise_distances(protein_xyz, ligand_xyz) Loading deepchem/feat/splif_fingerprints.py +1 −1 Original line number Diff line number Diff line Loading @@ -56,7 +56,7 @@ def compute_splif_features_in_range(frag1, } return (splif_dict) def featurize_splif(frag1, frag1, contact_bins, def featurize_splif(frag1, frag2, contact_bins, pairwise_distances, ecfp_degree): """Computes SPLIF featurization of fragment interactions binding pocket. Loading deepchem/feat/tests/test_contact_fingerprints.py +11 −0 Original line number Diff line number Diff line Loading @@ -31,3 +31,14 @@ class TestContactFeaturizers(unittest.TestCase): self.complex_files) assert features.shape == (1, voxels_per_edge, voxels_per_edge, voxels_per_edge, size) def test_contact_voxels_flattened(self): box_width = 48 voxel_width = 2 voxels_per_edge = box_width/voxel_width size = 8 voxelizer = dc.feat.ContactCircularVoxelizer(box_width=box_width, voxel_width=voxel_width, size=size, flatten=True) features, failures = voxelizer.featurize_complexes( self.complex_files) assert features.shape == (1, int(size*voxels_per_edge**3)) Loading
deepchem/feat/contact_fingerprints.py +17 −6 Original line number Diff line number Diff line Loading @@ -133,8 +133,10 @@ class ContactCircularVoxelizer(ComplexFeaturizer): in which they originated. Featurizes a macromolecular complex into a tensor of shape `(voxels_per_edge, voxels_per_edge, voxels_per_edge, size)` where `voxels_per_edge = int(box_width/voxel_width)`. `(voxels_per_edge, voxels_per_edge, voxels_per_edge, size)` where `voxels_per_edge = int(box_width/voxel_width)`. If `flatten==True`, then returns a flattened version of this tensor of length `size*voxels_per_edge**3` """ def __init__(self, Loading @@ -142,7 +144,8 @@ class ContactCircularVoxelizer(ComplexFeaturizer): radius=2, size=8, box_width=16.0, voxel_width=1.0): voxel_width=1.0, flatten=False): """ Parameters ---------- Loading @@ -157,6 +160,9 @@ class ContactCircularVoxelizer(ComplexFeaturizer): is centered on a ligand centroid. voxel_width: float, optional (default 1.0) Size of a 3D voxel in a grid. flatten: bool, optional (default False) If True, then returns a flat feature vector rather than voxel grid. This feature vector is constructed by flattening the usual voxel grid. """ self.cutoff = cutoff self.radius = radius Loading @@ -164,6 +170,7 @@ class ContactCircularVoxelizer(ComplexFeaturizer): self.box_width = box_width self.voxel_width = voxel_width self.voxels_per_edge = int(self.box_width / self.voxel_width) self.flatten = flatten def _featurize_complex(self, molecular_complex): """ Loading Loading @@ -193,7 +200,6 @@ class ContactCircularVoxelizer(ComplexFeaturizer): sum([ voxelize( convert_atom_to_voxel, self.voxels_per_edge, self.box_width, self.voxel_width, hash_ecfp, Loading @@ -209,5 +215,10 @@ class ContactCircularVoxelizer(ComplexFeaturizer): ecfp_degree=self.radius)) ]) ) # Features are of shape (voxels_per_edge, voxels_per_edge, voxels_per_edge, num_feat) so we should concatenate on the last axis. if self.flatten: return np.concatenate([features.flatten() for features in pairwise_features]) else: # Features are of shape (voxels_per_edge, voxels_per_edge, # voxels_per_edge, num_feat) so we should concatenate on the last # axis. return np.concatenate(pairwise_features, axis=-1)
deepchem/feat/rdkit_grid_featurizer.py +6 −8 Original line number Diff line number Diff line Loading @@ -37,8 +37,6 @@ logger = logging.getLogger(__name__) http://stackoverflow.com/questions/2827393/angles-between-two-n-dimensional-vectors-in-python """ # TODO(rbharath): Consider this comment on rdkit forums https://github.com/rdkit/rdkit/issues/1590 about sybyl featurization FLAT_FEATURES = ['ecfp_ligand', 'ecfp_hashed', 'splif_hashed', 'hbond_count'] VOXEL_FEATURES = [ Loading Loading @@ -306,12 +304,12 @@ class RdkitGridFeaturizer(ComplexFeaturizer): logger.warning("Some molecules cannot be loaded by Rdkit. Skipping") return None time1 = time.time() centroid = compute_centroid(ligand_xyz) ligand_xyz = subtract_centroid(ligand_xyz, centroid) protein_xyz = subtract_centroid(protein_xyz, centroid) time2 = time.time() logger.info("TIMING: Centroid processing took %0.3f s" % (time2 - time1)) #time1 = time.time() #centroid = compute_centroid(ligand_xyz) #ligand_xyz = subtract_centroid(ligand_xyz, centroid) #protein_xyz = subtract_centroid(protein_xyz, centroid) #time2 = time.time() #logger.info("TIMING: Centroid processing took %0.3f s" % (time2 - time1)) pairwise_distances = compute_pairwise_distances(protein_xyz, ligand_xyz) Loading
deepchem/feat/splif_fingerprints.py +1 −1 Original line number Diff line number Diff line Loading @@ -56,7 +56,7 @@ def compute_splif_features_in_range(frag1, } return (splif_dict) def featurize_splif(frag1, frag1, contact_bins, def featurize_splif(frag1, frag2, contact_bins, pairwise_distances, ecfp_degree): """Computes SPLIF featurization of fragment interactions binding pocket. Loading
deepchem/feat/tests/test_contact_fingerprints.py +11 −0 Original line number Diff line number Diff line Loading @@ -31,3 +31,14 @@ class TestContactFeaturizers(unittest.TestCase): self.complex_files) assert features.shape == (1, voxels_per_edge, voxels_per_edge, voxels_per_edge, size) def test_contact_voxels_flattened(self): box_width = 48 voxel_width = 2 voxels_per_edge = box_width/voxel_width size = 8 voxelizer = dc.feat.ContactCircularVoxelizer(box_width=box_width, voxel_width=voxel_width, size=size, flatten=True) features, failures = voxelizer.featurize_complexes( self.complex_files) assert features.shape == (1, int(size*voxels_per_edge**3))
