Frequently Asked Questions
This page summarizes practical input, output, visualization, and citation guidance for the pyDelPhi web server.
For best results, use a carefully prepared structure with appropriate protonation states, partial atomic charges,
and atomic radii.
Question: What is a .pqr file?
Answer:
A
.pqr file is an ASCII text structure file that contains atomic position coordinates,
partial atomic charges, and atomic radii. In simple terms, the name reflects:
- P: atomic position coordinates
- Q: partial atomic charge
- R: atomic radius or size
pyDelPhi requires this charge and radius information to solve the electrostatic problem. A coordinate-only
structure is usually not sufficient unless charges and radii are assigned by the server or by your own
preprocessing workflow.
Question: Should I upload a PQR file or a PDB file?
Answer:
PQR input is recommended because it already contains the charges and radii needed by pyDelPhi.
Uploading a PQR file gives you the most control over protonation state, ligand parameters, charge model, and
atomic radii.
PDB input is also supported, but a PDB file contains coordinates only. If you upload a PDB file, the server
must assign charge and size parameters using the selected force-field option. This is convenient, but it may
not be appropriate for every protonation state, ligand, modified residue, metal site, or nonstandard atom.
Question: What happens when I upload a PDB file?
Answer:
When a PDB file is uploaded, the server performs a basic automated preparation workflow. It uses pdb4amber to
clean the PDB structure and Reduce to add hydrogens and optimize local hydrogen-bonding choices, including
common side-chain flips and histidine tautomer choices.
This workflow is useful for many standard protein structures, but it is not equivalent to pKa prediction or full
pH-dependent protonation-state assignment. Tools such as H++, PROPKA, DelPhiPKa, or related workflows are more
appropriate when residue-specific protonation states at a target pH are important.
Therefore, PDB upload is convenient for standard structures, while a user-curated PQR file remains the recommended
input when pH effects, buried titratable residues, ligands, metals, modified residues, or charge/radius choices
are scientifically important.
Question: Does pyDelPhi accept a PQR file with missing hydrogen atoms?
Answer:
No. A PQR file should contain the hydrogens appropriate for the protonation state you intend to model.
Missing hydrogens can produce incorrect charges, incorrect local electrostatics, and unreliable results.
Prepare and inspect the PQR file before submission.
Question: What should I do if my structure has missing residues or missing atoms?
Answer:
Missing residues or missing atoms should be modeled or repaired before generating the final PQR file.
pyDelPhi is an electrostatics solver and should not be used as a structure-repair tool. If the missing region
is far from the region of interest, you may decide whether it is acceptable for your scientific use case, but
the server will not automatically rebuild it.
For protonation-state prediction and PQR preparation, you may use external tools or workflows, including
DelPhiPKa, PDB2PQR, AmberTools/Reduce, or other structure-preparation software appropriate for your system.
Question: Should I keep water, ions, ligands, or small molecules in my input structure?
Answer:
It depends on the scientific question. pyDelPhi can include atoms if they are present with appropriate coordinates,
charges, and radii. For PQR input, this means every retained atom, including ligand or ion atoms, must have valid
charge and radius values.
For PDB input, nonstandard ligands, modified residues, unusual ions, and custom molecules may not be safely
parameterized by default force-field choices. If these components are important, a curated PQR file is usually
the better input option.
Question: How should I handle small molecules bound to a protein?
Answer:
pyDelPhi can include small molecules if their atoms have valid coordinates, partial charges, and radii.
The recommended approach is to prepare a complete PQR file that includes the protein and the small molecule
with the charge/radius model you want to use.
If you upload a PDB file containing a ligand, verify that the selected parameter option supports that ligand.
Otherwise, prepare a custom PQR file or use a custom parameter workflow before submission.
Question: Can some atomic radii be zero?
Answer:
No. A zero atomic radius is unphysical for this calculation and can cause invalid or unreliable results.
Every atom included in the PQR file should have a positive, meaningful radius.
Question: Can all partial atomic charges be zero?
Answer:
No. If all charges are zero, the PQR file effectively lacks charge information. Partial atomic charge information
is required for meaningful electrostatic calculations. Some individual atoms may have zero charge depending on
the parameter set, but the entire molecule should not be uncharged unless that is a deliberate and justified
special case.
Question: Why does pyDelPhi prefer PQR instead of coordinate-only formats such as PDB?
Answer:
PDB files usually contain atomic coordinates but do not contain the electrostatic information needed by pyDelPhi.
PQR files include both coordinates and electrostatic parameters. This allows users to prepare the molecule for
the pH, protonation state, ligand model, charge model, and radius set that are appropriate for their study.
In other words, PQR input makes the chemical assumptions explicit.
Question: How can I generate a PQR file from a PDB file?
Answer:
Several tools can prepare PQR files or help assign protonation states and hydrogens. Options include:
- DelPhiPKa:
http://compbio.clemson.edu/pka_webserver/
- PDB2PQR/APBS workflows
- AmberTools/Reduce or related structure-preparation workflows
- Other molecular modeling tools that can assign protonation states, charges, and radii
After generating the PQR file, inspect it before submission. Confirm that hydrogens, charges, radii, ligands,
and any biologically important ions are represented as intended.
Question: When should I use custom charge and size files?
Answer:
Use custom charge and size files when the default parameter sets are not appropriate for your system.
This may include modified residues, unusual cofactors, custom ligands, nonstandard atom names, or a charge/radius
model chosen for a particular study.
Custom parameters should be used carefully. The atom naming and residue naming in your structure should match
the entries in the custom files. If they do not match, charges or radii may be missing or assigned incorrectly.
Question: What input file formats does the server accept?
Answer:
Uploaded files must be plain-text scientific input files in the supported formats listed below.
.pqr — structure file containing atom coordinates, partial charges, and atomic radii.pdb — coordinate file; the server performs basic cleanup and hydrogen addition before parameter assignment.crg — charge parameter file used for custom charge assignment.siz — atomic radius/size parameter file used for custom radius assignment
PQR file
A
.pqr file contains one atom per coordinate record. The final numeric fields provide the
partial atomic charge and atomic radius used by pyDelPhi.
Recommended PQR-style column layout:
Columns 1-6 Record name, usually ATOM or HETATM
Columns 7-11 Atom serial number
Columns 13-16 Atom name
Columns 18-20 Residue name
Column 22 Chain identifier
Columns 23-26 Residue number
Columns 31-38 X coordinate
Columns 39-46 Y coordinate
Columns 47-54 Z coordinate
Columns 55-62 Partial atomic charge
Columns 63-70 Atomic radius
Columns 78-80 Element symbol, when available
- Hydrogens should be present for the intended protonation state.
- Each atom should have a valid partial charge.
- Each atom should have a positive atomic radius.
- Atom and residue names should be consistent with the preparation workflow or parameter set used.
Example PQR excerpt
ATOM 1 N ALA A 1 11.104 13.207 8.678 -0.3000 1.8240 N
ATOM 2 H ALA A 1 10.550 13.815 9.233 0.3300 0.6000 H
ATOM 3 CA ALA A 1 12.532 13.470 8.820 0.2100 1.9080 C
ATOM 4 HA ALA A 1 12.973 12.929 9.650 0.1000 1.1000 H
ATOM 5 C ALA A 1 13.218 13.038 7.530 0.5100 1.9080 C
CRG file
A
.crg file defines partial charges for atom/residue names. It is used when custom charge
parameters are selected. Atom and residue naming should match the uploaded structure or a supported alias
in the selected parameter set.
Fixed-column CRG layout:
Columns 1-4 Atom name
Columns 7-9 Residue name
Column 11 Chain identifier, optional
Columns 12-15 Residue number, optional
Columns 16-end Charge value
Lines beginning with
!,
#, or
ATOM are treated as comments. Inline
comments beginning with
! or
# are also ignored.
- Each meaningful entry should identify an atom/residue name and a charge value.
- Charges should be numeric values such as
-0.500, 0.000, or 1.000.
- Use consistent atom and residue names across the structure, charge file, and size file.
Example CRG excerpt
!atom res chain resn charge
N ALA A 1 -0.3000
H ALA A 1 0.3300
CA ALA A 1 0.2100
HA ALA A 1 0.1000
C ALA A 1 0.5100
SIZ file
A
.siz file defines atomic radii/sizes for atom/residue names. It is used when custom size
parameters are selected. Atom and residue naming should match the uploaded structure or a supported alias
in the selected parameter set.
Fixed-column SIZ layout:
Columns 1-4 Atom name
Columns 7-9 Residue name
Column 11 Chain identifier, optional
Columns 13-end Radius/size value
Lines beginning with
!,
#, or
ATOM are treated as comments. Inline
comments beginning with
! or
# are also ignored.
- Each meaningful entry should identify an atom/residue name and a radius value.
- Radius values must be numeric and positive.
- A radius of
0.0 should not be used for atoms included in the calculation.
Example SIZ excerpt
!atom res chain size
N ALA A 1.8240
H ALA A 0.6000
CA ALA A 1.9080
HA ALA A 1.1000
C ALA A 1.9080
Allowed characters for CRG and SIZ files
Custom
.crg and
.siz files may contain only the following characters:
A-Z a-z 0-9 _ - + # ! . E e spaces tabs newlines carriage returns
In regular-expression form:
\A[A-Za-z0-9_\-+#!.Ee\t\r\n ]*\z
Practical recommendation
Prepare custom
.crg and
.siz files using a plain-text editor. Keep naming consistent
across the structure, charge file, and size file. For custom parameters, fixed-column alignment is recommended.
Question: What output files can I download?
Answer:
Finished jobs provide a summary report and a ZIP archive containing the available result files. Depending on
the options selected and the calculation settings, outputs may include:
run_summary.rpt — calculation summary and selected log informationoutputs.csv — tabulated energy or prediction values when availablepotential-map.cube, potential-map.phi, or potential-map.phi32 — electrostatic potential mapssurface-map.cube, surface-map.phi, or surface-map.phi32 — surface maps when requested/generatedzeta-potential.zphi — zeta-potential map when requested/generatedresult-files.zip — packaged result files
Dielectric/epsilon map output is currently not exposed as a regular form option because epsilon is used in a
more complex way internally and is not yet presented as a simple scalar 3D visualization map.
Question: How does interactive visualization work on the results page?
Answer:
If a finished job contains a supported structure file and one or more supported cube maps, the results page may
show an interactive visualization panel. The structure is loaded first, and map data is loaded lazily only when
you choose a map and click Load Visualization.
This design avoids loading large map files unnecessarily. It also keeps the page responsive for jobs where users
only want to download the output files.
Question: How long are my job results retained?
Answer:
To ensure optimal system performance, job files—including downloadable assets and interactive map
visualizations—are automatically cleared from the system after a fixed window (7 days) following their completion.
Once this period passes, the job status changes to "Expired" and the files cannot be recovered.
If you still need the results of an expired job, you can regenerate the data at any time by simply
resubmitting your original structure file and parameters.
Question: Do I need to cite the visualization software?
Answer:
If you use the interactive molecular visualization generated from this server in figures, presentations, or
publications, please acknowledge 3Dmol.js in addition to the appropriate DelPhi/pyDelPhi references.
Nicholas Rego and David Koes. 3Dmol.js: molecular visualization with WebGL.
Bioinformatics 31(8), 1322–1324, 2015.
doi:10.1093/bioinformatics/btu829
If you only download numerical DelPhi/pyDelPhi results and do not use the visualization, cite the relevant
DelPhi/pyDelPhi method and server references instead.
Question: How should I cite results from this server?
Answer:
Please cite the references most relevant to what you used. For most users, the primary citation is the
pyDelPhi calculation-engine paper, because pyDelPhi is the computational workhorse behind this server.
- If you use electrostatic calculation results from this server, please cite the pyDelPhi paper.
- If your work discusses DelPhi methodology, software lineage, or comparison with earlier DelPhi workflows, also cite the relevant DelPhi references.
- If your work specifically discusses the web-server workflow, cite the DelPhi web-server references where appropriate.
- If you use interactive visualization images generated from this server, also cite 3Dmol.js.
Primary citation: pyDelPhi calculation engine
Panday S.K., Zhao S., Alexov E. Accurate and Scalable Continuum Electrostatics
for Large Biomolecular Systems: The pyDelPhi Poisson–Boltzmann Framework.
Journal of Chemical Information and Modeling. 2025;66(1):488–502.
doi:10.1021/acs.jcim.5c02818.
DelPhi method and software lineage
Li C., Jia Z., Chakravorty A., Pahari S., Peng Y., Basu S., Koirala M.,
Panday S.K., Petukh M., Li L., Alexov E. DelPhi Suite: New Developments
and Review of Functionalities. Journal of Computational Chemistry.
2019;40(28):2502–2508. doi:10.1002/jcc.26006.
DelPhi web-server lineage
Sarkar S., Witham S., Zhang J., Zhenirovskyy M., Rocchia W., Alexov E.
DelPhi Web Server: A comprehensive online suite for electrostatic calculations
of biological macromolecules and their complexes.
Communications in Computational Physics. 2013;13:269–284.
Interactive visualization
Rego N., Koes D. 3Dmol.js: molecular visualization with WebGL.
Bioinformatics. 2015;31(8):1322–1324.
doi:10.1093/bioinformatics/btu829.
You do not need to cite every reference above for every use case. Cite pyDelPhi for calculations from this server,
add DelPhi lineage or web-server references when those are relevant to your discussion, and cite 3Dmol.js only when
you use visualization generated through the server.
Question: Why should I bookmark the results page?
Answer:
The results page URL contains the public task identifier needed to access the job. Bookmark or save the URL
after submission. The server does not expose a public directory listing of jobs, and the bookmark link is the
intended way to return to your result page.
Question: What should I check if my job fails?
Answer:
Common causes include malformed input files, missing charges, missing radii, missing hydrogens, unsupported atom
or residue names, inconsistent custom parameter files, or structures that require preparation before electrostatic
calculation.
Review the error report on the results page and inspect your input structure. If the issue persists, contact us
at
delphi@g.clemson.edu and include the task ID and a short description
of what you were trying to calculate.
