Protein analysis from scratch

This example treats proteins as structured coordinate data rather than just points in 3D space. It uses all three bundled PDB structures:

• 1fqy.pdb: backbone atoms, alpha carbons, contact maps and simplified secondary structure.
• 1aml.pdb: NMR ensemble contact frequency.
• 3ptb.pdb: ligands, waters and binding-site residues.

import molscope as ms

aqp = ms.read("examples/data/1fqy.pdb")
print(aqp.summary())
print(len(aqp.backbone()), "backbone atoms")
print(len(aqp.alpha_carbons()), "alpha carbons")

cmap = aqp.contact_map(cutoff=8.0, level="residue", method="ca", min_seq_sep=4)
print(cmap.n_contacts, "non-local CA contacts")

ss = aqp.secondary_structure()
print(ss.summary())

For an NMR ensemble:

models = ms.read_pdb_models("examples/data/1aml.pdb")
freq = ms.ensemble_contact_frequency(models, cutoff=8.0)
print(len(models), "models")
print(freq.matrix.shape, freq.n_contacts)

For a protein-ligand complex:

trypsin = ms.read("examples/data/3ptb.pdb")
print(trypsin.ligands())                 # [LigandResidue(A:BEN1, 9 atoms)]
print(len(trypsin.select(resname="HOH")), "water atoms")

site = trypsin.binding_site(cutoff=4.5)
for residue, distance in zip(site.residues[:8], site.min_distances[:8]):
    print(f"{residue!s:<10} {distance:.2f} A")

MolScope's secondary_structure() is a simplified, dependency-free DSSP-style assignment for teaching and prototyping. It is not a bit-identical replacement for canonical mkdssp.

Runnable versions:

uv run python examples/protein_analysis.py
uv run jupyter lab notebooks/protein_analysis_from_scratch.ipynb