File size: 8,103 Bytes
fb8a87c | 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 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 | from dataclasses import dataclass
from typing import Any, Sequence, TypeAlias, Union
import numpy as np
from .esmfold2_msa import MSA
# fmt: off
MSAInput: TypeAlias = Union[
MSA,
None,
]
# fmt: on
@dataclass
class Modification:
position: int # zero-indexed
ccd: str
smiles: str | None = None # TODO(mlee): add smiles support
@dataclass
class ProteinInput:
id: str | list[str]
sequence: str
modifications: list[Modification] | None = None
msa: MSAInput = None
@dataclass
class RNAInput:
id: str | list[str]
sequence: str
modifications: list[Modification] | None = None
@dataclass
class DNAInput:
id: str | list[str]
sequence: str
modifications: list[Modification] | None = None
@dataclass
class LigandInput:
id: str | list[str]
smiles: str | None = None
ccd: list[str] | None = None
@dataclass
class DistogramConditioning:
chain_id: str
distogram: np.ndarray
@dataclass
class PocketConditioning:
binder_chain_id: str
contacts: list[tuple[str, int]]
@dataclass
class CovalentBond:
chain_id1: str
res_idx1: int
atom_idx1: int
chain_id2: str
res_idx2: int
atom_idx2: int
@dataclass
class StructurePredictionInput:
sequences: Sequence[ProteinInput | RNAInput | DNAInput | LigandInput]
pocket: PocketConditioning | None = None
distogram_conditioning: list[DistogramConditioning] | None = None
covalent_bonds: list[CovalentBond] | None = None
def serialize_structure_prediction_input(all_atom_input: StructurePredictionInput):
def create_chain_data(seq_input, chain_type: str) -> dict[str, Any]:
chain_data: dict[str, Any] = {
"sequence": seq_input.sequence,
"id": seq_input.id,
"type": chain_type,
}
if hasattr(seq_input, "modifications") and seq_input.modifications:
mods = [
{"position": mod.position, "ccd": mod.ccd}
for mod in seq_input.modifications
]
chain_data["modifications"] = mods
if not hasattr(seq_input, "msa"):
pass
elif seq_input.msa is None:
chain_data["msa"] = None
elif isinstance(seq_input.msa, MSA):
chain_data["msa"] = {"sequences": seq_input.msa.sequences}
else:
error_msg = f"MSA must be None or MSA. Got {seq_input.msa} instead."
raise AttributeError(error_msg)
return chain_data
sequences = []
for seq_input in all_atom_input.sequences:
if isinstance(seq_input, ProteinInput):
sequences.append(create_chain_data(seq_input, "protein"))
elif isinstance(seq_input, RNAInput):
sequences.append(create_chain_data(seq_input, "rna"))
elif isinstance(seq_input, DNAInput):
sequences.append(create_chain_data(seq_input, "dna"))
elif isinstance(seq_input, LigandInput):
sequences.append(
{
"smiles": seq_input.smiles,
"id": seq_input.id,
"ccd": seq_input.ccd,
"type": "ligand",
}
)
else:
raise ValueError(f"Unsupported sequence input type: {type(seq_input)}")
result: dict[str, Any] = {"sequences": sequences}
if all_atom_input.covalent_bonds is not None:
result["covalent_bonds"] = [
{
"chain_id1": bond.chain_id1,
"res_idx1": bond.res_idx1,
"atom_idx1": bond.atom_idx1,
"chain_id2": bond.chain_id2,
"res_idx2": bond.res_idx2,
"atom_idx2": bond.atom_idx2,
}
for bond in all_atom_input.covalent_bonds
]
if all_atom_input.pocket is not None:
result["pocket"] = {
"binder_chain_id": all_atom_input.pocket.binder_chain_id,
"contacts": all_atom_input.pocket.contacts,
}
if all_atom_input.distogram_conditioning is not None:
result["distogram_conditioning"] = [
{"chain_id": disto.chain_id, "distogram": disto.distogram.tolist()}
for disto in all_atom_input.distogram_conditioning
]
return result
def deserialize_structure_prediction_input(
data: dict[str, Any],
) -> StructurePredictionInput:
"""Inverse of :func:`serialize_structure_prediction_input`.
Reconstructs a :class:`StructurePredictionInput` from the JSON-safe dict
produced by ``serialize_structure_prediction_input``. Values round-trip;
``DistogramConditioning.distogram`` dtype follows from JSON (``int64``
for integer entries, ``float64`` for floats) — cast back to the original
dtype if downstream code requires a specific one.
"""
def _mods(chain: dict[str, Any]) -> list[Modification] | None:
raw = chain.get("modifications")
if not raw:
return None
return [Modification(position=m["position"], ccd=m["ccd"]) for m in raw]
def _msa(chain: dict[str, Any]) -> MSAInput:
if "msa" not in chain or chain["msa"] is None:
return None
msa_blk = chain["msa"]
if isinstance(msa_blk, str):
raise ValueError(f"Unexpected MSA string value: {msa_blk!r}")
return MSA.from_sequences(msa_blk["sequences"])
sequences: list[ProteinInput | RNAInput | DNAInput | LigandInput] = []
for chain in data["sequences"]:
t = chain["type"]
if t == "protein":
sequences.append(
ProteinInput(
id=chain["id"],
sequence=chain["sequence"],
modifications=_mods(chain),
msa=_msa(chain),
)
)
elif t == "rna":
sequences.append(
RNAInput(
id=chain["id"],
sequence=chain["sequence"],
modifications=_mods(chain),
)
)
elif t == "dna":
sequences.append(
DNAInput(
id=chain["id"],
sequence=chain["sequence"],
modifications=_mods(chain),
)
)
elif t == "ligand":
sequences.append(
LigandInput(
id=chain["id"], smiles=chain.get("smiles"), ccd=chain.get("ccd")
)
)
else:
raise ValueError(f"Unsupported sequence type: {t!r}")
pocket: PocketConditioning | None = None
if (pocket_blk := data.get("pocket")) is not None:
pocket = PocketConditioning(
binder_chain_id=pocket_blk["binder_chain_id"],
contacts=[tuple(c) for c in pocket_blk["contacts"]],
)
distogram_conditioning: list[DistogramConditioning] | None = None
if (disto_blk := data.get("distogram_conditioning")) is not None:
distogram_conditioning = [
DistogramConditioning(
chain_id=d["chain_id"], distogram=np.asarray(d["distogram"])
)
for d in disto_blk
]
covalent_bonds: list[CovalentBond] | None = None
if (bonds_blk := data.get("covalent_bonds")) is not None:
covalent_bonds = [
CovalentBond(
chain_id1=b["chain_id1"],
res_idx1=b["res_idx1"],
atom_idx1=b["atom_idx1"],
chain_id2=b["chain_id2"],
res_idx2=b["res_idx2"],
atom_idx2=b["atom_idx2"],
)
for b in bonds_blk
]
return StructurePredictionInput(
sequences=sequences,
pocket=pocket,
distogram_conditioning=distogram_conditioning,
covalent_bonds=covalent_bonds,
)
|