prote

A Rust library for protein structure analysis.

What it does

• Parses mmCIF with its own lexer and parser.
• Builds a flat Structure (atoms, residues, entity instances, entities, bonds) and one Frame of positions per model and altloc.
• Bonds come from the file's _chem_comp_bond templates, plus backbone links between consecutive polymer residues (C–N for peptides, O3'–P for nucleic acids).
• Atom selections by instance, residue or atom.
• Spatial queries over a frame with a KD-tree.
• Contacts between two selections, excluding atoms within three bonds of each other.

Example

use std::fs;
use prote::{contact::Contact, mmcif::MmCIF, selection::Selection, spatial_index::SpatialIndex, structure::EntityKind, Cif};

let mmcif = MmCIF::new(Cif::new(&fs::read_to_string("3PTB.cif").unwrap())).unwrap();
let (structure, frames) = mmcif.build_structure().unwrap();
let index = SpatialIndex::new(&frames[0].positions);

let protein = Selection::by_instance(&structure, |_, entity| matches!(entity.kind, EntityKind::Polymer(_)));
let ligand = Selection::by_residue(&structure, |residue| residue.name == "BEN");

for contact in Contact::contacts(&structure, &index, &ligand, &protein) {
    let residue = &structure.residues[structure.residue_of(contact.atom_2)];
    println!("{} {} {:.2} Å", residue.name, residue.number, contact.distance);
}

examples/export_dot.rs writes Graphviz files of a structure's bonds and contacts.

Not yet

• Disulfides and other links from _struct_conn.
• Interaction types (H-bonds, salt bridges, metal coordination, hydrophobic, π-stacking).
• Rings.
• Structures whose file has no _chem_comp_bond (needs the Chemical Component Dictionary).

License

Copyright (c) 2026 Alan Halatian. Licensed under the GNU Affero General Public License v3.0 only; see LICENSE.