File size: 2,014 Bytes
ca7299e |
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 |
""" Metrics. """
import mdtraj as md
import numpy as np
from openfold.np import residue_constants
from tmtools import tm_align
from data import utils as du
def calc_tm_score(pos_1, pos_2, seq_1, seq_2):
tm_results = tm_align(pos_1, pos_2, seq_1, seq_2)
return tm_results.tm_norm_chain1, tm_results.tm_norm_chain2
def calc_mdtraj_metrics(pdb_path):
try:
traj = md.load(pdb_path)
pdb_ss = md.compute_dssp(traj, simplified=True)
pdb_coil_percent = np.mean(pdb_ss == 'C')
pdb_helix_percent = np.mean(pdb_ss == 'H')
pdb_strand_percent = np.mean(pdb_ss == 'E')
pdb_ss_percent = pdb_helix_percent + pdb_strand_percent
pdb_rg = md.compute_rg(traj)[0]
except IndexError as e:
print('Error in calc_mdtraj_metrics: {}'.format(e))
pdb_ss_percent = 0.0
pdb_coil_percent = 0.0
pdb_helix_percent = 0.0
pdb_strand_percent = 0.0
pdb_rg = 0.0
return {
'non_coil_percent': pdb_ss_percent,
'coil_percent': pdb_coil_percent,
'helix_percent': pdb_helix_percent,
'strand_percent': pdb_strand_percent,
'radius_of_gyration': pdb_rg,
}
def calc_aligned_rmsd(pos_1, pos_2):
aligned_pos_1 = du.rigid_transform_3D(pos_1, pos_2)[0]
return np.mean(np.linalg.norm(aligned_pos_1 - pos_2, axis=-1))
def calc_ca_ca_metrics(ca_pos, bond_tol=0.1, clash_tol=1.0):
ca_bond_dists = np.linalg.norm(
ca_pos - np.roll(ca_pos, 1, axis=0), axis=-1)[1:]
ca_ca_dev = np.mean(np.abs(ca_bond_dists - residue_constants.ca_ca))
ca_ca_valid = np.mean(ca_bond_dists < (residue_constants.ca_ca + bond_tol))
ca_ca_dists2d = np.linalg.norm(
ca_pos[:, None, :] - ca_pos[None, :, :], axis=-1)
inter_dists = ca_ca_dists2d[np.where(np.triu(ca_ca_dists2d, k=0) > 0)]
clashes = inter_dists < clash_tol
return {
'ca_ca_deviation': ca_ca_dev,
'ca_ca_valid_percent': ca_ca_valid,
'num_ca_ca_clashes': np.sum(clashes),
}
|