tools-only Structure Navigation - Usage Guide
Overview
This skill covers navigating the SMCRA (Structure-Model-Chain-Residue-Atom) hierarchy in Biopython Bio.PDB. Use it to access specific parts of structures, iterate over components, extract sequences, and handle disordered atoms.
Prerequisites
pip install biopython
Quick Start
Tell your AI agent what you want to do:
- "List all chains in this PDB file"
- "Extract the sequence from chain A"
- "Find all ligands in this structure"
Example Prompts
Accessing Structure Parts
"Get chain A from model 0"
"Show me all atoms in residue 100"
"What is the B-factor of the CA atom in residue 50?"
Iterating
"Print all residue numbers in chain B"
"List all C-alpha coordinates"
"Count atoms in each chain"
Extracting Information
"Extract the protein sequence from this structure"
"Find all ARG residues"
"List all hetero groups (ligands)"
Handling Disorder
"Show alternative conformations for residue 42"
"List all disordered atoms in chain A"
What the Agent Will Do
- Parse the structure file
- Navigate to requested level using SMCRA hierarchy
- Extract requested information (coordinates, properties, sequences)
- Handle disordered atoms/residues appropriately
- Return organized results
SMCRA Hierarchy
Structure (whole PDB entry)
Model (NMR conformer or asymmetric unit)
Chain (polypeptide or nucleic acid)
Residue (amino acid, nucleotide, ligand, water)
Atom (individual atom with coordinates)
Tips
- Use
get_methods for iteration:get_chains(),get_residues(),get_atoms() - Residue ID is a tuple -
(hetfield, resseq, icode), not just a number - hetfield values:
' '= amino acid,'W'= water,'H_xxx'= hetero - PPBuilder extracts polypeptide sequences from coordinates
- CaPPBuilder works better for structures with missing atoms
- Selection.unfold_entities extracts all entities at a level (R=residue, A=atom)