NEW: dynamic generation of SMILES for PTMs

Author: Mikhail Lebedev

alphabase/constants/const_files/smiles.yaml

@@ -0,0 +1,79 @@
+aa_smiles:
+  A: N([Xe])([Xe])[C@@]([H])(C)C(=O)[Rn]
+  R: N([Xe])([Xe])[C@@]([H])(CCCNC(=N)N)C(=O)[Rn]
+  N: N([Xe])([Xe])[C@@]([H])(CC(=O)N)C(=O)[Rn]
+  D: N([Xe])([Xe])[C@@]([H])(CC(=O)O)C(=O)[Rn]
+  C: N([Xe])([Xe])[C@@]([H])(CS)C(=O)[Rn]
+  Q: N([Xe])([Xe])[C@@]([H])(CCC(=O)N)C(=O)[Rn]
+  E: N([Xe])([Xe])[C@@]([H])(CCC(=O)O)C(=O)[Rn]
+  G: N([Xe])([Xe])CC(=O)[Rn]
+  H: N([Xe])([Xe])[C@@]([H])(CC1=CN=C-N1)C(=O)[Rn]
+  I: N([Xe])([Xe])[C@@]([H])([C@]([H])(CC)C)C(=O)[Rn]
+  L: N([Xe])([Xe])[C@@]([H])(CC(C)C)C(=O)[Rn]
+  K: N([Xe])([Xe])[C@@]([H])(CCCCN)C(=O)[Rn]
+  M: N([Xe])([Xe])[C@@]([H])(CCSC)C(=O)[Rn]
+  F: N([Xe])([Xe])[C@@]([H])(Cc1ccccc1)C(=O)[Rn]
+  P: N1([Xe])[C@@]([H])(CCC1)C(=O)[Rn]
+  S: N([Xe])([Xe])[C@@]([H])(CO)C(=O)[Rn]
+  T: N([Xe])([Xe])[C@@]([H])([C@]([H])(O)C)C(=O)[Rn]
+  W: N([Xe])([Xe])[C@@]([H])(CC(=CN2)C1=C2C=CC=C1)C(=O)[Rn]
+  Y: N([Xe])([Xe])[C@@]([H])(Cc1ccc(O)cc1)C(=O)[Rn]
+  V: N([Xe])([Xe])[C@@]([H])(C(C)C)C(=O)[Rn]
+
+n_term_modifications:
+  mTRAQ: C(=O)CN1CCN(CC1)C
+  mTRAQ:d4: C(=O)[13C]([H])([H])[15N]1[13C]([H])([H])[13C]([H])([H])N(CC1)C
+  mTRAQ:d8: C(=O)[13C]([H])([H])[15N]1[13C]([H])([H])[13C]([H])([H])[15N]([13C]([H])([H])[13C]1([H])([H]))[13C]([H])([H])([H])
+  Acetyl: C(=O)C
+  Propionyl: C(=O)CC
+  Biotin: C(=O)CCCCC1SCC2NC(=O)NC21
+  Carbamidomethyl: C(=O)NC
+  Carbamyl: C(=O)N
+  Propionamide: C(=O)CC(N)=O
+  Pyridylacetyl: C(=O)Cc1ccccn1
+  Methyl: C
+  Dimethyl: C
+  Dimethyl:2H(6)13C(2): '[13C]([2H])([2H])([2H])'
+  Dimethyl:2H(4): C([2H])([2H])([1H])
+  Dimethyl:2H(4)13C(2): '[13C]([2H])([2H])([1H])'
+
+c_term_modifications:
+  Methyl: OC
+
+ptm_dict:
+  Carbamidomethyl@C: C(C(C(=O)[Rn])N([Xe])([Xe]))SCC(=O)N
+  Oxidation@M: O=C([Rn])C(N([Xe])([Xe]))CCS(=O)C
+  GlyGly@K: NCC(=O)NCC(=O)NCCCC[C@H](N([Xe])([Xe]))C([Rn])=O
+  Deamidated@N: C([C@@H](C(=O)[Rn])N([Xe])([Xe]))C(=O)O
+  Propionyl@K: CCC(=O)NCCCCC(C(=O)[Rn])N([Xe])([Xe])
+  Deamidated@Q: C(CC(=O)O)[C@@H](C(=O)[Rn])N([Xe])([Xe])
+  Gln->pyro-Glu@Q^Any N-term: O=C([Rn])[C@H]1N([Xe])C(=O)CC1
+  Glu->pyro-Glu@E^Any N-term: O=C([Rn])[C@H]1N([Xe])C(=O)CC1
+  Phospho@S: O=P(O)(O)OC[C@@H](C(=O)[Rn])N([Xe])([Xe])
+  Nitro@Y: O=[N+]([O-])c1cc(ccc1O)C[C@@H](C(=O)[Rn])N([Xe])([Xe])
+  Acetyl@K: CC(=O)NCCCC[C@H](N([Xe])([Xe]))C(=O)[Rn]
+  Dimethyl@K: CN(C)CCCC[C@H](N([Xe])([Xe]))C(=O)[Rn]
+  mTRAQ@K: '[H]N(CCCC[C@H](N([Xe])([Xe]))C(=O)[Rn])C(=O)CN1CCN(C)CC1'
+  mTRAQ:d4@K: '[H]N(CCCC[C@H](N([Xe])([Xe]))C(=O)[Rn])C(=O)[13CH2][15N]1CCN(C)[13CH2][13CH2]1'
+  mTRAQ:d8@K: '[H]N(CCCC[C@H](N([Xe])([Xe]))C(=O)[Rn])C(=O)[13CH2][15N]1[13CH2][13CH2][15N]([13CH3])[13CH2][13CH2]1'
+  Pyridylethyl@C: C1=CN=CC=C1CCSCC(C(=O)[Rn])N([Xe])([Xe])
+  Butyryl@K: CCCC(=O)NCCCCC(C(=O)[Rn])N([Xe])([Xe])
+  Phospho@T: CC(C(C(=O)[Rn])N([Xe])([Xe]))OP(=O)(O)O
+  Methylthio@C: CSSC[C@H](N([Xe])([Xe]))C([Rn])=O
+  Carbamidomethyl@M: C[S+](CCC(N([Xe])([Xe]))C([Rn])=O)CC(N)=O
+  Succinyl@K: C(CCN)CC(C(=O)[Rn])N([Xe])C(=O)CCC(=O)O
+  Crotonyl@K: CC=CC(=O)NCCCCC(C(=O)[Rn])N([Xe])([Xe])
+  Phospho@Y: C1=CC(=CC=C1CC(C(=O)[Rn])N([Xe])([Xe]))OP(=O)(O)O
+  Malonyl@K: N([Xe])([Xe])[C@@H](CCCC(NC(=O)CC(=O)O)=O)C(=O)[Rn]
+  Met->Hse@M: N([Xe])([Xe])[C@H](C(=O)[Rn])CCO
+  Pro->(2S,4R)-4-fluoroproline@P: F[C@@H]1C[C@H](N([Xe])C1)C(=O)[Rn]
+  Pro->(2S, 4S)-4fluoroproline@P: F[C@H]1C[C@H](N([Xe])C1)C(=O)[Rn]
+  Pro->(2S)-1,3-thiazolidine-2-carboxylic acid@P: S1[C@H](N([Xe])CC1)C(=O)[Rn]
+  Pro->(4R)-1,3-Thiazolidine-4-carboxylic acid@P: S1CN([Xe])[C@@H](C1)C(=O)[Rn]
+  Pro->(2S,4R)-4-hydroxyproline@P: O[C@@H]1C[C@H](N([Xe])C1)C(=O)[Rn]
+  Pro->(DL)-pipecolic acid@P: C1CCN([Xe])C(C1)C(=O)[Rn]
+  Pro->3,4-Dehydro-L-proline@P: C1C=CC(N1([Xe]))C(=O)[Rn]
+  Pro->(1S,3S,5S)-2-Azabicyclo[3.1.0]hexane-3-carboxylic acid@P: '[C@H]12N([Xe])[C@@H](C[C@@H]2C1)C(=O)[Rn]'
+  Pro->(1R,3S,5R)-2-Azabicyclo[3.1.0]hexane-3-carboxylic acid@P: '[C@@H]12N([Xe])[C@@H](C[C@H]2C1)C(=O)[Rn]'
+  Pro->(2S,3aS,7aS)-Octahydro-1H-indole-2-carboxylic acid@P: N1([Xe])[C@@H](C[C@@H]2CCCC[C@H]12)C(=O)[Rn]
+  Pro->(DL)-5-trifluoromethylproline@P: FC(C1CCC(N1([Xe]))C(=O)[Rn])(F)F

alphabase/smiles/smiles.py

@@ -0,0 +1,99 @@
+from typing import Optional
+import os
+from rdkit import Chem
+from rdkit.Chem.rdmolops import ReplaceSubstructs, SanitizeMol
+
+from alphabase.constants._const import CONST_FILE_FOLDER
+from alphabase.yaml_utils import load_yaml
+
+SMILES_ALL: dict = load_yaml(os.path.join(CONST_FILE_FOLDER, "smiles.yaml"))
+
+N_TERM_PLACEHOLDER = "[Xe]"
+C_TERM_PLACEHOLDER = "[Rn]"
+
+aa_smiles = SMILES_ALL["aa_smiles"]
+n_term_modifications = SMILES_ALL["n_term_modifications"]
+c_term_modifications = SMILES_ALL["c_term_modifications"]
+ptm_dict = SMILES_ALL["ptm_dict"]
+
+aa_mols = {aa: Chem.MolFromSmiles(smile) for aa, smile in aa_smiles.items()}
+
+
+def validate_correct_args(
+    aa_smiles: str, n_term_mod: Optional[str] = None, c_term_mod: Optional[str] = None
+) -> None:
+    """
+    Validate the amino acid, N-terminal modification, and C-terminal modification.
+
+    Args:
+        aa_smiles (str): SMILES of an amino acid with or without PTMs, has to have the placeholder atoms [Xe] for N-term and [Rn] for C-term.
+        n_term_mod (Optional[str]): N-terminal modification name. Options are the keys in the n_term_modifications dict.
+        c_term_mod (Optional[str]): C-terminal modification name. Options are the keys in the c_term_modifications dict.
+
+    Raises:
+        ValueError: If the amino acid is invalid, or if N-terminal / C-terminal modification is not recognized.
+    """
+    if N_TERM_PLACEHOLDER not in aa_smiles or C_TERM_PLACEHOLDER not in aa_smiles:
+        raise ValueError(
+            f"The amino acid {aa_smiles} must contain the N-terminal and C-terminal placeholders in order to be modified."
+        )
+    if n_term_mod is not None and n_term_mod not in n_term_modifications:
+        raise ValueError(f"Unrecognized N-terminal modification: {n_term_mod}")
+    if c_term_mod is not None and c_term_mod not in c_term_modifications:
+        raise ValueError(f"Unrecognized C-terminal modification: {c_term_mod}")
+    _mol = Chem.MolFromSmiles(aa_smiles)
+    if _mol is None:
+        raise ValueError(f"Invalid amino acid SMILES: {aa_smiles}")
+
+
+def modify_amino_acid(
+    aa_smiles: str, n_term_mod: Optional[str] = None, c_term_mod: Optional[str] = None
+) -> str:
+    """
+    Modify an amino acid with N-terminal and C-terminal modifications.
+
+    Args:
+        aa (str): SMILES of an amino acid with or without PTMs, has to have the placeholder atoms [Xe] for N-term and [Rn] for C-term.
+        n_term_mod (Optional[str]): N-terminal modification name. Options are the keys in the n_term_modifications dict. If None, the amino acid is not modified, so using hydrogen atoms.
+        c_term_mod (Optional[str]): C-terminal modification name. Options are the keys in the c_term_modifications dict. If None, the amino acid is not modified, so using the -OH group.
+
+    Returns:
+        str: SMILES string of the modified amino acid.
+    """
+    validate_correct_args(aa_smiles, n_term_mod, c_term_mod)
+    mol = Chem.MolFromSmiles(aa_smiles)
+    n_term_placeholder_mol = Chem.MolFromSmiles(N_TERM_PLACEHOLDER, sanitize=False)
+    c_term_placeholder_mol = Chem.MolFromSmiles(C_TERM_PLACEHOLDER, sanitize=False)
+
+    # Apply N-terminal modification
+    if n_term_mod:
+        n_mod_smiles = n_term_modifications[n_term_mod]
+        n_mod_mol = Chem.MolFromSmiles(n_mod_smiles)
+        mol = ReplaceSubstructs(mol, n_term_placeholder_mol, n_mod_mol)[0]
+
+        if "Dimethyl" in n_term_mod:
+            mol = ReplaceSubstructs(mol, n_term_placeholder_mol, n_mod_mol)[0]
+        else:
+            mol = ReplaceSubstructs(
+                mol, n_term_placeholder_mol, Chem.MolFromSmiles("[H]", sanitize=False)
+            )[0]
+    else:
+        mol = ReplaceSubstructs(
+            mol,
+            n_term_placeholder_mol,
+            Chem.MolFromSmiles("[H]", sanitize=False),
+            replaceAll=True,
+        )[0]
+
+    # Apply C-terminal modification
+    if c_term_mod:
+        c_mod_smiles = c_term_modifications[c_term_mod]
+        c_mod_mol = Chem.MolFromSmiles(c_mod_smiles)
+        mol = ReplaceSubstructs(mol, c_term_placeholder_mol, c_mod_mol)[0]
+    else:
+        mol = ReplaceSubstructs(
+            mol, c_term_placeholder_mol, Chem.MolFromSmiles("O", sanitize=False)
+        )[0]
+
+    SanitizeMol(mol)
+    return Chem.MolToSmiles(mol)