Atom layout in a fixed shape (usually 1-dim or 2-dim). Examples for atom layouts are atom37, atom14, and similar. All members are np.ndarrays wit

Items of _chem_comp category. For the full list of items and their semantics see http://mmcif.rcsb.org/dictionaries/mmcif_pdbx_v50.dic/Categories

Self attention that is either per-sequence or per-residue.

Raised when the structure store doesn't contain the requested target.

Rigid Transformation, i.e. element of special euclidean group.

Structure class for representing and processing molecular structures.

Transition block for transformer.

Triangle Multiplication.

Extra information for chemical components occurring in mmCIF. Fields: chem_comp: A mapping from _chem_comp.id to associated items in the

Gather indices to translate from one atom layout to another. All members are np or jnp ndarray (usually 1-dim or 2-dim) with the same shape, e.g.

Container for alphabet specific (dna, rna, protein) atom information.

Failed alignment between the hit sequence and the actual mmCIF sequence.

Table of atoms in a Structure.

Table of chains in a Structure.

Error indicating a problem with MSA Search.

An error indicating that template is invalid.

Ligand input. Attributes: id: Unique ligand "chain" identifier. ccd_ids: The Chemical Component Dictionary or user-defined CCD IDs of the

Dataclass containing MSA.

Protein chain input.

Error reading a record.

List of residues with meta data. Attributes: res_name: np.ndarray of str [num_res], e.g. 'ARG', 'TRP' res_id: np.ndarray of int [num_res]

Table of residues in a Structure.

RNA chain input.

Rot3Array Matrix in 3 dimensional Space implemented as struct of arrays.

A container for templates that were found for the given query sequence. The structure_store is constructed from the config by default. Callers can

Raised if total number of residues for all chains outside allowed range.

A mapping from internal values to author values in a mmCIF. Fields: auth_asym_id: A mapping from label_asym_id to auth_asym_id. auth_seq_id

Stores and processes bioassembly data from mmCIF tables.

Describes a bond between two atoms.

Exception raised by errors when getting bond atom indices.

Table of atomic bonds.

Items of _chem_comp_atom category. See mmcif.wwpdb.org/dictionaries/mmcif_ma.dic/Categories/chem_comp_atom.html Attributes: type_symbol: _ch

Items of _chem_comp_bond category. See mmcif.wwpdb.org/dictionaries/mmcif_ma.dic/Categories/_chem_comp_bond.html Attributes: atom_id_1: _che

Single strand DNA chain input.

Template hit metrics derived from the MSA for filtering and featurising. Attributes: pdb_id: The PDB ID of the hit. auth_chain_id: The auth

An error indicating that invalid release date was detected.

Postprocessed model result. Attributes: predicted_structure: Predicted protein structure. numerical_data: Useful numerical data (scalars or

MSA Attention.

Error raised when an atom is missing during alignment.

Raised when bioassembly data is missing from an mmCIF.

Raised when chemical components data is missing from an mmCIF.

Raised if the mmcif file contains too many / too few chains.

Helper class to run structure prediction stages.

Raised if the data pipeline produces data containing nans.

Raise when the mmCIF does not have any atoms.

A rigid transformation operation.

Computed mean outer product.

Stores attributes of a processed inference result. Attributes: cif: CIF file containing an inference result. mean_confidence_1d: Mean 1D co

View of a table with renamed column names.

Atom-level data arrays with a residue dimension. Attributes: atom_positions: float32 of shape [num_res, num_atom_type, 3] coordinates. atom

Stores the inference results (diffusion samples) for a single seed. Attributes: seed: The seed used to generate the samples. inference_resu

Embed a single template.

Structural template input.

Dataclass containing templates.

Dataclass containing features for tokens.

A rigid transformation composed of a sequence of `Operation`s.

Vec3Array in 3 dimensional Space implemented as struct of arrays. This is done in order to improve performance and precision. On TPU small matrix

Class for incrementally building chain and residue tables.

Helper class for slicing the leading dimensions of a `Structure`. Wraps a `Structure` instance and applies a slice operation to the coordinate fi

A file-like object that presents a concatenated view of multiple files.

A pickle `Unpickler` that forbids loading arbitrary global classes.

Dataclass for 1D per-atom confidences from AlphaFold.

Operate on flat atoms.

Config base class. Subclassing Config automatically makes the subclass a kw_only dataclass with a `__post_init__` that coerces Config-subclass fi

Dataclass containing batch.

Chemical Components found in PDB (CCD) constants. See https://academic.oup.com/bioinformatics/article/31/8/1274/212200 for CCD CIF format documen

Confidence categories for AlphaFold predictions.

Head to predict the distance errors in a prediction.

Configuration for Evoformer.

Config for PairFormerIteration.

Configuration for EvoformerIteration.

Configuration for ConfidenceHead.

Configuration for DiffusionHead.

Configuration object for `WholePdbPipeline`. Properties: max_atoms_per_token: number of atom slots in one token (was called num_den

Metaclass that synthesizes a __post_init__ that coerces dicts to Config subclass instances.