PDB2PQR

What is PDB2PQR?

PDB2PQR prepares biomolecular structures for electrostatics calculations by converting PDB or mmCIF files to PQR format. The PQR format extends PDB by replacing the occupancy and B-factor columns with atomic partial charge and radius values—the two parameters required for solving the Poisson-Boltzmann equation.

Continuum electrostatics methods like Poisson-Boltzmann calculations reveal how charge distribution influences protein stability, binding affinity, and molecular recognition. However, PDB files from X-ray crystallography lack hydrogen atoms and contain no charge or radius information. PDB2PQR automates the preparation pipeline: reconstructing missing atoms, predicting protonation states at your target pH, and assigning force field parameters.

The output is a PQR file ready for electrostatics solvers like APBS (Adaptive Poisson-Boltzmann Solver). For general structure repair before molecular dynamics, consider PDB Fixer instead—it offers more extensive modeling capabilities including loop reconstruction and solvent box construction.

How does PDB2PQR work?

PDB2PQR operates as a multi-step pipeline that analyzes, corrects, and parameterizes protein structures.

Debumping

Steric clashes between residues are resolved by systematically rotating dihedral angles. The algorithm uses distance cutoffs of 1.0 Å for hydrogen-hydrogen collisions, 1.5 Å for hydrogen-heavy atom interactions, and 2.0 Å for heavy atom pairs. When conflicts are detected, dihedral angles are adjusted in 5-degree increments until a non-clashing configuration is found.

Hydrogen optimization

Missing hydrogen atoms are added and then optimized to improve the hydrogen bonding network. The optimization includes:

• Side chain flipping for HIS, ASN, and GLN residues to optimize polar contacts
• Rotational adjustment of hydroxyl hydrogens on SER, THR, TYR, and CYS
• Optimal placement of protons on neutral HIS, protonated GLU, and protonated ASP
• Water hydrogen reorientation throughout the structure

Protonation state assignment with PROPKA

PROPKA predicts pKa values for all titratable residues based on the local protein environment—nearby charges, hydrogen bonds, and solvent accessibility shift pKa values from their reference values. At your specified pH, residues are assigned appropriate protonation states.

For example, a buried glutamate forming a salt bridge might have its pKa shifted upward, remaining protonated (neutral) at physiological pH where surface glutamates would be deprotonated (charged).

Charge and radius assignment

After determining the structure and protonation states, PDB2PQR assigns atomic partial charges ($q$) and radii ($r$) from the selected force field. These parameters define the inputs to the Poisson-Boltzmann equation:

$$\nabla \cdot [\varepsilon(\mathbf{r}) \nabla \phi(\mathbf{r})] - \bar{\kappa}^2(\mathbf{r}) \sinh[\phi(\mathbf{r})] = -\frac{4\pi \rho(\mathbf{r})}{k_B T}$$

Where $\varepsilon(\mathbf{r})$ is the dielectric function (determined by atomic radii), $\phi(\mathbf{r})$ is the electrostatic potential, and $\rho(\mathbf{r})$ is the charge density (from atomic charges).

Inputs & settings

Input requirements

Upload a PDB (.pdb or .ent) or mmCIF (.cif or .mmcif) protein, DNA, or RNA structure, or fetch a structure directly from RCSB using a PDB ID. Submitted structure bytes and native format are preserved for PDB2PQR. PDB2PQR cannot model large missing regions, so ensure the structure is reasonably complete before processing.

You can optionally upload a MOL2 file to parameterize one ligand. The ligand name must match exactly one ligand in each submitted structure. Custom force fields require both the native DAT parameter file and names file.

Force field selection

The force field determines which charge and radius parameters are assigned to each atom. This choice significantly affects your electrostatics results.

• AMBER: General-purpose force field widely used in molecular dynamics. Choose this when comparing PB results with explicit solvent AMBER simulations.
• CHARMM: Another MD-focused force field. Preferred when your downstream analysis uses CHARMM-based tools.
• PARSE: Specifically optimized for Poisson-Boltzmann calculations. We recommend PARSE for most electrostatics applications including surface visualization, binding energy calculations, and pKa prediction.
• PEOEPB: Partial Equalization of Orbital Electronegativities optimized for PB calculations. An alternative to PARSE for non-standard chemistries.
• SWANSON: AMBER ff99 charges with radii optimized for PB calculations. Combines AMBER compatibility with PB-optimized radii.
• TYL06: Another PB-optimized parameter set based on careful fitting to experimental solvation energies.

PARSE is the recommended choice for standalone electrostatics calculations. Use AMBER or CHARMM when you need consistency with MD simulation parameters.

Protonation settings

• Use PROPKA for protonation: Enables pKa prediction to determine protonation states. Disabled by default to match native PDB2PQR behavior; when enabled, standard protonation states are adjusted based on pH.
• pH value: The pH for protonation state assignment. Common values: 7.0 for physiological conditions, 4.5 for endosomal/lysosomal compartments, 2.0 for gastric conditions.
• pH analysis: Choose a single pH or an evenly spaced series of 2–7 points. A series performs a complete, independent PDB2PQR run at every pH and returns every PQR file; it does not interpolate charges or reuse protonation states between points.
• PROPKA advanced options: Limit titration handling to selected residues, supply a custom parameter file, and include coupled-residue diagnostics.

Structure options

• Remove water molecules: Strips crystallographic water from the output. Enable for electrostatics surface visualization; disable to preserve waters that may influence local electrostatics.
• Keep chain IDs: Preserves original chain identifiers when needed. Disabled by default to match the stock PDB2PQR CLI output.
• Neutral termini: Uses PDB2PQR’s native neutral N- or C-terminus modes. These modes require the PARSE force field.

Advanced settings

• Debump structure: Controls the native steric-clash removal pass. Disable to pass --nodebump.
• Optimize hydrogens: Controls the native hydrogen-bond optimization step. Disable to pass --noopt.
• Assign charges only: Uses --assign-only to assign charges and radii without atom rebuilding, debumping, or hydrogen optimization. Use it only with an already complete, explicitly protonated structure.
• Output naming scheme: Uses --ffout so residue and atom names follow a selected force-field naming convention.
• Include original header: Preserves the uploaded structure header in the generated PQR remarks.
• Whitespace-delimited PQR: Uses PDB2PQR’s native spacing option for programs that prefer separated coordinate fields.
• Generate PDB output: Downloads the optional native PDB-format output alongside the main PQR file.
• Generate APBS input: Downloads the optional APBS .in template generated from the final PQR file.
• Run APBS electrostatics: Runs APBS 3.4.1 using PDB2PQR’s generated input and returns the native OpenDX potential map, APBS input, and APBS diagnostics. This option already includes the APBS input, so it cannot be combined with the input-only switch.
• Run log: Returns PDB2PQR’s native .log diagnostics alongside each successful result.
• Clean only: Reformats and cleans the PDB without adding atoms or assigning parameters. Use this when you only need structure cleanup without force field assignment.

Understanding the results

PDB2PQR outputs a PQR file containing atomic coordinates, charges, and radii. The results summary shows structure statistics:

For CIF/mmCIF inputs or whitespace-delimited PQR output, residue and chain counts are unavailable unless Keep chain IDs is enabled. In those modes, native PDB2PQR can omit both chain identifiers and TER separators, so the output no longer contains enough information for an accurate count.

When PROPKA is enabled, the pKa values table exposes the native summary rows in their original order, including residue/group identity, chain, predicted pKa, model pKa, and applied pH. The Run coverage table pairs PDB2PQR’s own parsed input counts with output counts, total charge, native warning messages, and APBS status. These diagnostics report what the programs emitted; they do not invent a parameterization score or silently classify unsupported chemistry as successful.

Every successful native run also includes a JSON reproducibility manifest with the exact runtime versions, normalized command arguments, applied pH, molecule kind, and SHA-256 digests for the input and returned artifacts. Temporary service paths are excluded.

Viewing APBS electrostatics

With Run APBS electrostatics enabled, the Electrostatics result displays the prepared PQR structure together with the native OpenDX potential. Red and blue isosurfaces show −1 and +1 kT/e, respectively. The PQR, OpenDX, APBS input, and APBS log remain independently downloadable for analysis in PyMOL, VMD, or another compatible program.

Interpreting total charge

The total charge should match your expectations based on the amino acid composition and pH:

• At neutral pH, expect charge ≈ (Arg + Lys) − (Asp + Glu) plus contributions from the termini
• Net charge affects electrostatic potential maps and can indicate protonation state issues if unexpected
• Highly positive or negative values may indicate pH is far from pI

Using the PQR output

The PQR file can be used directly with APBS for calculating:

• Electrostatic potential surfaces for visualization and analysis
• Binding free energies using PB continuum methods
• pKa shifts when comparing different conformational states
• Solvation energies for thermodynamic analysis

Best practices

Use PARSE for pure electrostatics work. This force field was designed specifically for Poisson-Boltzmann calculations and provides the most accurate results for solvation energies and potential surfaces.

Match your pH to experimental conditions. If studying enzyme activity at pH 5, use pH 5.0 for protonation—the charge distribution at different pH values can dramatically alter electrostatic properties.

Inspect unusual total charges. If the net charge seems unexpected, examine the PROPKA output for residues with shifted pKa values. A buried aspartate might remain protonated when you expected it to be charged.

Keep crystallographic waters for binding studies. Waters at interfaces often mediate important electrostatic interactions. Remove them only when computing surface potentials for visualization.

Common workflows

Electrostatic surface visualization

Prepare the structure with PARSE and PROPKA at your pH of interest, then enable Run APBS electrostatics. The result includes the interactive electrostatic view and the native files needed for PyMOL, VMD, or downstream automation.

Comparing protonation across pH

Enable PROPKA, choose pH series, and set the endpoints and number of points. Compare the pKa and coverage tables by applied pH, inspect total-charge changes, and download each independently prepared PQR. Enabling APBS also produces one potential map per permitted series run.

Comparing wild-type and mutant electrostatics

Process both structures with identical settings (same force field, pH, options). The difference in electrostatic potential reveals how mutations alter the charge distribution and local environment—useful for understanding effects on binding or stability.

Structure preparation pipeline

For structures that need both repair and electrostatics preparation, run PDB Fixer first to add missing residues and atoms, then process the fixed structure through PDB2PQR for charge assignment.

Limitations

PDB2PQR works with protein and nucleic-acid structures but has limited support for non-standard molecules. Ligands, cofactors, and modified residues may not be parameterized in the standard force fields.

Some protonation states predicted by PROPKA may not be supported by all force fields. PARSE supports the widest range of protonation states including neutral termini, while AMBER and CHARMM have more limited options.

PDB2PQR jobs are limited to 10 structure–pH combinations. APBS jobs are limited to 3 combinations because each calculation generates a three-dimensional grid, and an individual OpenDX map must be no larger than 160 MiB for safe interactive delivery. These limits are checked before dispatch; reduce the number of structures or pH points when a request exceeds them.

The debumping algorithm finds the first acceptable configuration rather than the optimal one. For structures with severe clashes, additional energy minimization may be needed.