Bio Files: Read and write common biology file formats

This Rust and Python library contains functionality to load and save data in common biology file formats. It operates on data structures that are specific to each file format; you will need to convert to and from the structures used by your application. The API docs, and examples below are sufficient to get started.

Note: Install the pip version with pip install biology-files due to a name conflict.

Supported formats:

• mmCIF (Protein atom, residue, chain, and related data like secondary structure)
• Mol2 (Small molecules, e.g. ligands)
• SDF (Small molecules, e.g. ligands)
• PDBQT (Small molecules, e.g. ligands. Includes docking-specific fields.)
• Map (Electron density, e.g. from crystallography, Cryo EM)
• AB1 (Sequence tracing)
• DAT (Amber force field data for small molecules)
• FRCMOD (Amber force field patch data for small molecules)
• Amber .lib files, e.g. with charge data for amino acids and proteins.

Planned:

• MTZ (Exists in Daedalus; needs to be decoupled)

• DNA (Exists in PlasCAD; needs to be decoupled)

• CIF structure formats (2fo-fc etc) (Exists in Daedalus; needs to be decoupled)

For Genbank, we recommend gb-io. We do not plan to support this format, due to this high quality library.

Each module represents a file format, and most have dedicated structs dedicated to operating on that format.

It operates using structs with public fields, which you can explore using the API docs, or your IDE. These structs generally include these three methods: new(), save() and load(). new() accepts &str for text files, and a R: Read + Seek for binary. save() and load() accept &Path. The Force Field formats use load_dat, save_frcmod instead, as they use the same structs for both formats.

Serial numbers

Serial numbers for atoms, residues, secondary structure, and chains are generally pulled directly from atom data files (mmCIF, Mol2 etc). These lists reference atoms, or residues, stored as Vec<u32>, with the u32 being the serial number. In your application, you may wish to adapt these generic types to custom ones that use index lookups instead of serial numbers. We use SNs here because they're more robust, and match the input files directly; add optimizations downstream, like converting to indices, and/or applying back-references. (e.g. the index of the residue an atom's in, in your derived Atom struct).

Example use

Small molecule save and load, Python.

from biology_files import Sdf

sdf_data = Sdf.load("./molecules/DB03496.sdf")

sdf_data.atoms[0]
#AtomGeneric { serial_number: 1, posit: Vec3 { x: 2.3974, y: 1.1259, z: 2.5289 }, element: Chlorine, 
type_in_res: None, force_field_type: None, occupancy: None, partial_charge: None, hetero: true }

>>> sdf_data.atoms[0].posit
# [2.3974, 1.1259, 2.5289]

sdf_data.save("test.sdf")

mol2_data = sdf_data.to_mol2()
mol2_data.save("test.mol2")

Small molecule save and load, Rust.

use bio_files::{Sdf, Mol2};}

// ...
let sdf_data = Sdf::load("./molecules/DB03496.sdf");

sdf_data.atoms[0]; // (as above)
sdf_data.atoms[0].posit;  // (as above, but lin_alg::Vec3))

sdf_data.save("test.sdf");

let mol2_data: Mol2 = sdf_data.into();
mol2_data.save("test.mol2");


// Loading Force field parameters:
let p = Path::new("gaff2.dat")
let params = ForceFieldParams::load_dat(p)?;

You can use similar syntax for mmCIF protein files.

Amber force fields

Reference the Amber 2025 Reference Manual, section 15 for details on how we parse its files, and how to use the results. In some cases, we change the format from the raw Amber data. For example, we store angles as radians (vice degrees), and σ vice R_min for Van der Waals parameters. Structs and fields are documented with reference manual references.

The Amber forcefield parameter format has fields which each contain a Vec of a certain type of data. (Bond stretching parameters, angle between 3 atoms, torsion/dihedral angles etc.) You may wish to parse these into a format that has faster lookups for your application.

Note that the above examples expect that your application has a struct representing the molecule that has From<Mol2>, and to_mol2(&self) (etc) methods. The details of these depend on the application. For example:

impl From<Sdf> for Molecule {
    fn from(m: Sdf) -> Self {
        // We've implemented `From<AtomGeneric>` and `From<ResidueGeneric>` for our application's `Atom` and
        // `Residue`
        let atoms = m.atoms.iter().map(|a| a.into()).collect();
        let residues = m.residues.iter().map(|r| r.into()).collect();

        Self::new(m.ident, atoms, m.chains.clone(), residues, None, None);
    }
}

A practical example of parsing a molecule from a mmCIF as parsed from bio_files into an application-specific format:

fn load() {
    let cif_data = mmcif::load("./1htm.cif");
    let mol: Molecule = cif_data.try_into().unwrap();
}

impl TryFrom<MmCif> for Molecule {
    type Error = io::Error;

    fn try_from(m: MmCif) -> Result<Self, Self::Error> {
        let mut atoms: Vec<_> = m.atoms.iter().map(|a| a.into()).collect();

        let mut residues = Vec::with_capacity(m.residues.len());
        for res in &m.residues {
            residues.push(Residue::from_generic(res, &atoms)?);
        }

        let mut chains = Vec::with_capacity(m.chains.len());
        for c in &m.chains {
            chains.push(Chain::from_generic(c, &atoms, &residues)?);
        }

        // Now that chains and residues are loaded, update atoms with their back-ref index.
        for atom in &mut atoms {
            for (i, res) in residues.iter().enumerate() {
                if res.atom_sns.contains(&atom.serial_number) {
                    atom.residue = Some(i);
                    break;
                }
            }

            for (i, chain) in chains.iter().enumerate() {
                if chain.atom_sns.contains(&atom.serial_number) {
                    atom.chain = Some(i);
                    break;
                }
            }
        }

        let mut result = Self::new(m.ident.clone(), atoms, chains, residues, None, None);

        result.experimental_method = m.experimental_method.clone();
        result.secondary_structure = m.secondary_structure.clone();

        result.bonds_hydrogen = Vec::new();
        result.adjacency_list = result.build_adjacency_list();

        Ok(result)
    }
}

A protein loading and prep example:

Python:

use biology_files::{Mol2, MmCif, ForceFieldParams, FfParamSet, prepare_peptide, load_prmtop};

mol = Mol2.load("CPB.mol2")
protein = MmCif.load("1c8k.cif")

param_set = FfParamSet.new_amber()
lig_specific = ForceFieldParams.load_frcmod("CPB.frcmod")

# Or, instead of loading atoms and mol-specific params separately:
# mol, lig_specific = load_prmtop("my_mol.prmtop")

# Add Hydrogens, force field type, and partial charge to atoms in the protein; these usually aren't
# included from RSCB PDB. You can also call `populate_hydrogens_dihedrals()`, and
# `populate_peptide_ff_and_q() separately. Add bonds.
protein.atoms, protein.bonds = prepare_peptide(
    protein.atoms,
    protein.bonds,
    protein.residues,
    protein.chains,
    param_set.peptide_ff_q_map,
    7.0,
)

Rust:

use bio_files::{MmCif, Mol2, ForceFieldParams, FfParamSet, prepare_peptide, load_prmtop};
use std::path::Path;

fn load() {
    let param_set = FfParamSet::new_amber().unwrap();

    let mut protein = MmCif::load(Path::new("1c8k.cif")).unwrap();
    let mol = Mol2::load(Path::new("CPB.mol2")).unwrap();
    let mol_specific = ForceFieldParams::load_frcmod(Path::new("CPB.frcmod")).unwrap();

    // Or, instead of loading atoms and mol-specific params separately:
    // let (mol, lig_specific) = load_prmtop("my_mol.prmtop");

    // Or, if you have a small molecule available in Amber Geostd, load it remotely:
    // let data = bio_apis::amber_geostd::load_mol_files("CPB");
    // let mol = Mol2::new(&data.mol2);
    // let mol_specific = ForceFieldParams::from_frcmod(&data.frcmod);

    // Add Hydrogens, force field type, and partial charge to atoms in the protein; these usually aren't
    // included from RSCB PDB. You can also call `populate_hydrogens_dihedrals()`, and
    // `populate_peptide_ff_and_q() separately. Add bonds.
    prepare_peptide(
        &mut protein.atoms,
        &mut protein.bonds,
        &mut protein.residues,
        &mut protein.chains,
        &param_set.peptide_ff_q_map.as_ref().unwrap(),
        7.0,
    )
        .unwrap();
}

Note: The Python version is currently missing support for some formats, and not all fields are exposed.

References

• Amber 2025 Reference Manual, section 15