get losses

get losses

prep inputs for binder design --------------------------------------------------- -binder_len = length of binder to hallucinate (option i

prep inputs for fixed backbone design --------------------------------------------------- if copies > 1: -homooligomer=True - input

prep inputs for hallucination --------------------------------------------------- if copies > 1: -repeat=True - tie the repeating s

prep input for partial hallucination --------------------------------------------------- -length=100 - total length of protein (if differ

Worker function for parallel Rosetta relax. Runs in a child process with its own PyRosetta initialization. On exception, writes a full traceb

Worker function for AbMPNN sequence generation in multiprocessing. On exception, dumps the full traceback to ``{output_path}.err`` so the

Apply inverse Rigid3Array transform to point.

Return a new QuatAffine with tensor_fn applied to the rotation part.

Return a new QuatAffine with tensor_fn applied (e.g. stop_gradient).

Check that array and Matrix match.

Convert atom14 to atom37 representation.

Convert atom14 to atom37 representation.

Convert atom14 to atom37 representation.

Convert Atom37 positions to Atom14 positions.

Computes the frames for the up to 8 rigid groups for each residue. The rigid groups are defined by the possible torsions in a given amino acid. W

Computes the frames for the up to 8 rigid groups for each residue.

Computes the 7 torsion angles (in sin, cos encoding) for each residue. The 7 torsion angles are in the order '[pre_omega, phi, psi, chi_1, chi_2,

Backbone FAPE Loss.

Backbone FAPE Loss. Jumper et al. (2021) Suppl. Alg. 20 "StructureModule" line 17 Args: ret: Dictionary to write outputs into, needs to cont

Implements a JAX equivalent of `tf.gather` with `axis` and `batch_dims`.

Flat-bottom loss to penalize structural violations between residues. This is a loss penalizing any violation of the geometry around the peptide b

Flat-bottom loss to penalize structural violations between residues.

Loss to penalize steric clashes between residues. This is a loss penalizing any steric clashes due to non bonded atoms in different peptides comi

Loss to penalize steric clashes between residues.

Creates float32 variables when bfloat16 is requested.

Casts float32 to bfloat16 when bfloat16 was originally requested.

One block of the template pair stack.

Calculate secondary structure percentages with interface analysis. Analyzes protein secondary structure using DSSP and calculates percentages

Calculate Local Interaction Score (LIS) for protein complexes. Args: pdb_path: Path to PDB file pae_matrix: PAE matrix (nump

Find atom14 positions, this includes finding the correct renaming.

Computes the chi angles given all atom positions and the amino acid type. Args: positions: A Vec3Array of shape [num_res, residue_constan

Compute Frames from all atom positions. Args: aatype: array of aatypes, int of [N] all_atom_positions: Vector of all atom positions, shape

Computes per-residue pLDDT from logits. Args: logits: [num_res, num_bins] output from the PredictedLDDTHead. Returns: plddt: [num_res] per

Computes aligned confidence metrics from logits. Args: logits: [num_res, num_res, num_bins] the logits output from PredictedAlignedErrorHe

Find optimal renaming of ground truth based on the predicted positions. Jumper et al. (2021) Suppl. Alg. 26 "renameSymmetricGroundTruthAtoms" Th

Compute several metrics to assess the structural violations.

Compute several metrics to assess the structural violations.

compute agreement between predicted and measured contact map

Converts MSA in Stockholm format to the A3M format.

Create a deep copy of a dictionary with all nested structures. Performs deep copying of a dictionary to ensure complete independence from

Computes torsion angle for a quadruple of points. For points (a, b, c, d), this is the angle between the planes defined by points (a, b, c) and (

Applies module + dropout + residual update.

Computes euclidean distance between 'vec1' and 'vec2'. Args: vec1: Vec3Array to compute euclidean distance to vec2: Vec3Array to compute eu

Counts residues whose Ca is a large distance from its neighbour. Measures the fraction of CA-CA pairs between consecutive amino acids that are mo

Counts residues whose Ca is a large distance from its neighbor.