Merge pull request #9 from cmungall/program-learning-claude

Program learning claude

learned/claude-3-sonnet/11_12_saturated_fatty_acyl_CoA_4__.py

@@ -0,0 +1,55 @@
+"""
+Classifies: CHEBI:84948 11,12-saturated fatty acyl-CoA(4-)
+"""
+"""
+Classifies: CHEBI:84948 11,12-saturated fatty acyl-CoA(4-)
+"""
+from rdkit import Chem
+from rdkit.Chem import AllChem
+from rdkit.Chem import rdMolDescriptors
+
+def is_11_12_saturated_fatty_acyl_CoA_4__(smiles: str):
+    """
+    Determines if a molecule is a 11,12-saturated fatty acyl-CoA(4-) based on its SMILES string.
+    A 11,12-saturated fatty acyl-CoA(4-) is a fatty acyl-CoA(4-) where the 11-12 bond of the fatty acyl group is saturated.
+
+    Args:
+        smiles (str): SMILES string of the molecule
+
+    Returns:
+        bool: True if molecule is a 11,12-saturated fatty acyl-CoA(4-), False otherwise
+        str: Reason for classification
+    """
+
+    # Parse SMILES
+    mol = Chem.MolFromSmiles(smiles)
+    if mol is None:
+        return False, "Invalid SMILES string"
+
+    # Check for CoA substructure
+    coa_pattern = Chem.MolFromSmarts("C(C)(COP(=O)([O-])OP(=O)([O-])OC[C@H]1O[C@H]([C@H](O)[C@@H]1OP(=O)([O-])[O-])n1cnc2c(N)ncnc12)C(=O)NCCC(=O)NCCS")
+    if not mol.HasSubstructMatch(coa_pattern):
+        return False, "No CoA substructure found"
+
+    # Check for fatty acyl chain
+    acyl_pattern = Chem.MolFromSmarts("[CX3](=[OX1])[CX4,CX3]~[CX4,CX3]~[CX4,CX3]~[CX4,CX3]")
+    acyl_matches = mol.GetSubstructMatches(acyl_pattern)
+    if not acyl_matches:
+        return False, "No fatty acyl chain found"
+
+    # Check for saturation at 11-12 position
+    for acyl_match in acyl_matches:
+        acyl_atoms = [mol.GetAtomWithIdx(idx) for idx in acyl_match]
+        if len(acyl_atoms) < 12:
+            continue  # Acyl chain too short
+
+        atom11 = acyl_atoms[10]
+        atom12 = acyl_atoms[11]
+        if atom11.GetIsAromatic() or atom12.GetIsAromatic():
+            continue  # Aromatic atoms
+
+        bond = mol.GetBondBetweenAtoms(atom11.GetIdx(), atom12.GetIdx())
+        if bond.GetBondType() == Chem.BondType.SINGLE:
+            return True, "Fatty acyl chain has a saturated bond at position 11-12"
+
+    return False, "Fatty acyl chain does not have a saturated bond at position 11-12"

learned/claude-3-sonnet/11_oxo_steroid.py

@@ -0,0 +1,56 @@
+"""
+Classifies: CHEBI:47787 11-oxo steroid
+"""
+"""
+Classifies: 11-oxo steroid
+Definition: Any oxo steroid that has an oxo substituent at position 11
+"""
+from rdkit import Chem
+from rdkit.Chem import AllChem
+
+def is_11_oxo_steroid(smiles: str):
+    """
+    Determines if a molecule is an 11-oxo steroid based on its SMILES string.
+    
+    Args:
+        smiles (str): SMILES string of the molecule
+        
+    Returns:
+        tuple: (bool, str) - (True if molecule is an 11-oxo steroid, reason for classification)
+    """
+    # Parse SMILES
+    mol = Chem.MolFromSmiles(smiles)
+    if mol is None:
+        return False, "Invalid SMILES string"
+
+    # Check for basic steroid core (four fused rings)
+    steroid_core = Chem.MolFromSmarts("[#6]1~[#6]~[#6]~[#6]2~[#6]~[#6]~[#6]~3~[#6]~[#6]~[#6]~4~[#6]~[#6]~[#6]~[#6]~[#6]~4~[#6]~3~[#6]~2~[#6]~1")
+    if not mol.HasSubstructMatch(steroid_core):
+        return False, "No steroid core structure found"
+
+    # SMARTS pattern for 11-oxo group in steroid context
+    # This pattern looks for the specific environment of the 11-position ketone
+    # in the steroid ring system
+    oxo_11_pattern = Chem.MolFromSmarts("[#6]1~[#6]~[#6]2~[#6]~[#6]~[#6]~3~[#6](=[O:1])~[#6]~[#6]~[#6]~[#6]~3~[#6]~2~[#6]~[#6]~1")
+
+    # Find matches for the 11-oxo pattern
+    matches = mol.GetSubstructMatches(oxo_11_pattern)
+    if not matches:
+        return False, "No ketone group at position 11"
+
+    # Count carbons to verify it's in the typical steroid range
+    carbon_count = len([atom for atom in mol.GetAtoms() if atom.GetAtomicNum() == 6])
+    if carbon_count < 19 or carbon_count > 30:
+        return False, f"Carbon count ({carbon_count}) outside typical steroid range (19-30)"
+
+    # Additional check for reasonable molecular weight
+    mol_wt = Chem.Descriptors.ExactMolWt(mol)
+    if mol_wt < 250 or mol_wt > 500:
+        return False, f"Molecular weight ({mol_wt:.1f}) outside typical steroid range (250-500)"
+
+    # Count rings to ensure we have the right ring system
+    ring_info = mol.GetRingInfo()
+    if ring_info.NumRings() < 4:
+        return False, "Insufficient number of rings for steroid structure"
+
+    return True, "Molecule contains steroid core with ketone group at position 11"

learned/claude-3-sonnet/11beta_hydroxy_steroid.py

@@ -0,0 +1,60 @@
+"""
+Classifies: CHEBI:35346 11beta-hydroxy steroid
+"""
+"""
+Classifies: 11beta-hydroxy steroids
+"""
+from rdkit import Chem
+from rdkit.Chem import AllChem
+
+def is_11beta_hydroxy_steroid(smiles: str):
+    """
+    Determines if a molecule is an 11beta-hydroxy steroid based on its SMILES string.
+    
+    Args:
+        smiles (str): SMILES string of the molecule
+        
+    Returns:
+        bool: True if molecule is an 11beta-hydroxy steroid, False otherwise
+        str: Reason for classification
+    """
+
+    # Parse SMILES
+    mol = Chem.MolFromSmiles(smiles)
+    if mol is None:
+        return False, "Invalid SMILES string"
+
+    # Check for basic steroid core (4 fused rings)
+    steroid_core = Chem.MolFromSmarts("[#6]~1~[#6]~[#6]~[#6]~2~[#6]~[#6]~[#6]~[#6]~3~[#6]~[#6]~[#6]~[#6]~4~[#6]~[#6]~[#6]~[#6]~4~[#6]~3~[#6]~2~1")
+    if not mol.HasSubstructMatch(steroid_core):
+        return False, "No steroid core structure found"
+
+    # Check for 11-beta hydroxy group
+    # The SMARTS pattern looks for:
+    # - The specific carbon at position 11 in the steroid skeleton
+    # - An OH group in beta configuration (specified by '@H')
+    # Note: The exact stereochemistry is critical here
+    hydroxy_11beta = Chem.MolFromSmarts("[C]~1~[C]~[C]~[C]~2~[C]~[C]~[C]~[C]~3~[C]~[C@@H](O)~[C]~[C]~4~[C]~[C]~[C]~[C]~4~[C]~3~[C]~2~1")
+
+    if not mol.HasSubstructMatch(hydroxy_11beta):
+        return False, "No 11-beta hydroxy group found"
+
+    # Additional check for common substituents often found in 11beta-hydroxy steroids
+    # Look for common functional groups like ketones, other hydroxyls, etc.
+    ketone_pattern = Chem.MolFromSmarts("C(=O)")
+    other_hydroxy = Chem.MolFromSmarts("[CH]O")
+
+    if not (mol.HasSubstructMatch(ketone_pattern) or mol.HasSubstructMatch(other_hydroxy)):
+        return False, "Missing typical steroid substituents"
+
+    # Count carbons to ensure reasonable size for a steroid
+    carbon_count = len([atom for atom in mol.GetAtoms() if atom.GetAtomicNum() == 6])
+    if carbon_count < 19 or carbon_count > 30:  # Most steroids have 19-30 carbons
+        return False, f"Carbon count ({carbon_count}) outside typical steroid range"
+
+    # Check for reasonable number of rings
+    ring_info = mol.GetRingInfo()
+    if ring_info.NumRings() < 4:
+        return False, "Too few rings for steroid structure"
+
+    return True, "Contains steroid core with 11-beta hydroxy group and appropriate substituents"

learned/claude-3-sonnet/16beta_hydroxy_steroid.py

@@ -0,0 +1,61 @@
+"""
+Classifies: CHEBI:17354 16beta-hydroxy steroid
+"""
+"""
+Classifies: CHEBI:16beta-hydroxy steroid
+A 16-hydroxy steroid in which the hydroxy group at position 16 has a beta-configuration.
+"""
+from rdkit import Chem
+from rdkit.Chem import AllChem
+
+def is_16beta_hydroxy_steroid(smiles: str):
+    """
+    Determines if a molecule is a 16beta-hydroxy steroid based on its SMILES string.
+    
+    Args:
+        smiles (str): SMILES string of the molecule
+        
+    Returns:
+        bool: True if molecule is a 16beta-hydroxy steroid, False otherwise
+        str: Reason for classification
+    """
+    # Parse SMILES
+    mol = Chem.MolFromSmiles(smiles)
+    if mol is None:
+        return False, "Invalid SMILES string"
+
+    # Add explicit hydrogens for stereochemistry
+    mol = Chem.AddHs(mol)
+
+    # Basic steroid core pattern (four fused rings)
+    # Using SMARTS that matches the basic steroid skeleton
+    steroid_core = Chem.MolFromSmarts("[#6]~1~[#6]~[#6]~[#6]~2~[#6]~[#6]~[#6]~3~[#6]~[#6]~[#6]~4~[#6]~[#6]~[#6]~[#6]~4~[#6]~3~[#6]~2~1")
+
+    if not mol.HasSubstructMatch(steroid_core):
+        return False, "No steroid core structure found"
+
+    # Pattern for 16-beta-hydroxy group
+    # The [H] specifies explicit hydrogen, OH is the hydroxy group
+    # The @ symbols specify the stereochemistry
+    beta_oh_pattern = Chem.MolFromSmarts("[C]12[C][C@H](O)[CH2][C@]1([CH2,CH3])[C@@H]3[C][C][C]2")
+
+    if not mol.HasSubstructMatch(beta_oh_pattern):
+        return False, "No 16-beta-hydroxy group found"
+
+    # Additional checks for reasonable molecular weight and atom counts
+    # Steroids typically have molecular weights between 250-1000
+    mol_weight = Chem.Descriptors.ExactMolWt(mol)
+    if mol_weight < 250 or mol_weight > 1000:
+        return False, f"Molecular weight {mol_weight} outside typical steroid range (250-1000)"
+
+    # Count carbons (steroids typically have 17+ carbons)
+    carbon_count = len([atom for atom in mol.GetAtoms() if atom.GetAtomicNum() == 6])
+    if carbon_count < 17:
+        return False, f"Too few carbons ({carbon_count}) for a steroid structure"
+
+    # Check for at least one oxygen (for the hydroxy group)
+    oxygen_count = len([atom for atom in mol.GetAtoms() if atom.GetAtomicNum() == 8])
+    if oxygen_count < 1:
+        return False, "No oxygen atoms found"
+
+    return True, "Contains steroid core with 16-beta-hydroxy group"

learned/claude-3-sonnet/17alpha_hydroxy_steroid.py

@@ -0,0 +1,57 @@
+"""
+Classifies: CHEBI:35342 17alpha-hydroxy steroid
+"""
+"""
+Classifies: 17alpha-hydroxy steroid
+"""
+from rdkit import Chem
+from rdkit.Chem import AllChem
+
+def is_17alpha_hydroxy_steroid(smiles: str):
+    """
+    Determines if a molecule is a 17alpha-hydroxy steroid based on its SMILES string.
+    
+    Args:
+        smiles (str): SMILES string of the molecule
+        
+    Returns:
+        bool: True if molecule is a 17alpha-hydroxy steroid, False otherwise
+        str: Reason for classification
+    """
+
+    # Parse SMILES
+    mol = Chem.MolFromSmiles(smiles)
+    if mol is None:
+        return False, "Invalid SMILES string"
+
+    # Check for basic steroid core (4 fused rings)
+    steroid_core = Chem.MolFromSmarts("[#6]~1~[#6]~[#6]~[#6]~2~[#6]~[#6]~[#6]~[#6]~3~[#6]~[#6]~[#6]~[#6]~4~[#6]~[#6]~[#6]~[#6]~4~[#6]~3~[#6]~2~1")
+    if not mol.HasSubstructMatch(steroid_core):
+        return False, "No steroid core structure found"
+
+    # Check for 17-OH with alpha stereochemistry
+    # [C] is carbon 17, [OH1] is the hydroxyl group, '@' indicates stereochemistry
+    # The exact SMARTS pattern depends on the numbering convention used in the structure
+    oh_17_alpha_pattern = Chem.MolFromSmarts('[C;R1]-[C;R1]([OH1])')
+
+    if not mol.HasSubstructMatch(oh_17_alpha_pattern):
+        return False, "No hydroxyl group at C17 position found"
+
+    # Additional checks to verify it's a steroid structure:
+    # Count rings
+    ri = mol.GetRingInfo()
+    if ri.NumRings() < 4:
+        return False, "Insufficient number of rings for steroid structure"
+
+    # Basic size check - steroids typically have at least 19 carbons
+    c_count = sum(1 for atom in mol.GetAtoms() if atom.GetAtomicNum() == 6)
+    if c_count < 19:
+        return False, "Too few carbons for steroid structure"
+
+    # Check for sp3 hybridized carbons typical in steroid core
+    sp3_carbons = sum(1 for atom in mol.GetAtoms() 
+                     if atom.GetAtomicNum() == 6 and atom.GetHybridization() == Chem.HybridizationType.SP3)
+    if sp3_carbons < 10:
+        return False, "Insufficient sp3 carbons for steroid structure"
+
+    return True, "Contains steroid core with 17-alpha hydroxyl group"

learned/claude-3-sonnet/17beta_hydroxy_steroid.py

@@ -0,0 +1,70 @@
+"""
+Classifies: CHEBI:35343 17beta-hydroxy steroid
+"""
+"""
+Classifies: 17beta-hydroxy steroid
+A 17-hydroxy steroid in which the hydroxy group at position 17 has a beta-configuration.
+"""
+from rdkit import Chem
+from rdkit.Chem import AllChem
+
+def is_17beta_hydroxy_steroid(smiles: str):
+    """
+    Determines if a molecule is a 17beta-hydroxy steroid based on its SMILES string.
+    
+    Args:
+        smiles (str): SMILES string of the molecule
+        
+    Returns:
+        bool: True if molecule is a 17beta-hydroxy steroid, False otherwise
+        str: Reason for classification
+    """
+    # Parse SMILES
+    mol = Chem.MolFromSmiles(smiles)
+    if mol is None:
+        return False, "Invalid SMILES string"
+
+    # Check for basic steroid core (four fused rings)
+    steroid_core = Chem.MolFromSmarts("[#6]~1~[#6]~[#6]~[#6]~2~[#6]~[#6]~[#6]~[#6]~3~[#6]~[#6]~[#6]~[#6]~4~[#6]~[#6]~[#6]~[#6]~1~[#6]~2~[#6]~3~4")
+    if not mol.HasSubstructMatch(steroid_core):
+        return False, "No steroid core structure found"
+
+    # Check for 17-OH group in beta configuration
+    # [C] is carbon 17, [OH1] is hydroxy group, '@' indicates stereochemistry
+    # The [C] must be connected to 4 atoms (saturated)
+    # Note: The exact SMARTS pattern depends on the numbering convention used
+    oh_17_beta = Chem.MolFromSmarts('[C;X4](@[*])(@[*])(@[*])[OH1]')
+
+    if not mol.HasSubstructMatch(oh_17_beta):
+        return False, "No hydroxyl group with correct connectivity found"
+
+    # Get matches for OH group
+    oh_matches = mol.GetSubstructMatches(oh_17_beta)
+
+    # Check if any of the matches are at position 17
+    found_17_beta_oh = False
+    for match in oh_matches:
+        c_atom = mol.GetAtomWithIdx(match[0])  # Get the carbon atom
+        # Check if this carbon is part of the D ring (ring 4) of the steroid
+        # by checking its environment
+        ring_info = mol.GetRingInfo()
+        if ring_info.NumAtomRings(match[0]) > 0:  # Carbon must be part of a ring
+            # Check chirality of the carbon
+            if c_atom.GetChiralTag() == Chem.ChiralType.CHI_TETRAHEDRAL_CCW:
+                found_17_beta_oh = True
+                break
+
+    if not found_17_beta_oh:
+        return False, "No 17-beta hydroxyl group found"
+
+    # Additional validation: molecule should have reasonable size for a steroid
+    num_atoms = mol.GetNumAtoms()
+    if num_atoms < 20 or num_atoms > 100:
+        return False, "Molecule size not consistent with steroid structure"
+
+    # Count rings
+    ring_info = mol.GetRingInfo()
+    if ring_info.NumRings() < 4:
+        return False, "Insufficient number of rings for steroid structure"
+
+    return True, "Contains steroid core with 17-beta hydroxyl group"