Radius Configuration

FastSASA assigns a radius to every included atom before SASA calculation.

Assignment Order

1. User-provided --config-file FILE.
2. FASTSASA_DEFAULT_CONFIG, when set.
3. Bundled or installed share/protor.config.
4. Element-radius fallback for unknown residue/atom pairs.

When running from a source tree or an installed prefix, FastSASA searches common share/protor.config locations automatically. If you run the binary from an unusual directory layout, pass --config-file or set FASTSASA_DEFAULT_CONFIG so the intended radius table is unambiguous.

When FastSASA uses an element fallback, it prints a warning:

FastSASA: warning: atom 'NAG C1' unknown, guessing element is 'C', and radius 1.700 A

The atom remains included. Use --unknown skip or --unknown halt when guessed radii are not acceptable.

Configuration Format

A radius file contains a types: section and an atoms: section:

types:
C_ALIPHATIC 1.70 apolar
N_POLAR     1.55 polar
O_POLAR     1.52 polar

atoms:
ANY N  N_POLAR
ANY CA C_ALIPHATIC
ANY O  O_POLAR

The first section defines named radius classes. The second maps residue and atom names to those classes.

Use a custom file:

fastsasa --config-file my_radii.config structure.pdb

or for a trajectory:

fastsasa trajectory \
  --topology topology.pdb \
  --trajectory trajectory.xtc \
  --config-file my_radii.config \
  --filter protein

Nonstandard Residues

The bundled table covers standard residues and common structural entries. Unknown ligands, glycans, lipids, and modified residues fall back to element radii with warnings unless the user provides a reviewed configuration.

Do not treat element fallback as a force-field-specific parameterization. For publication work involving nonstandard chemistry, document the radius source and provide the configuration file.

Bundled ProtOr Table

FastSASA's default configuration uses the ProtOr atomic-group radii reported by Tsai, Taylor, Chothia, and Gerstein (1999). The table maps standard protein, nucleic-acid, and common cap atom names onto those published types. The original data attribution and redistribution notice are recorded in the repository NOTICE.

The element fallback is an approximation for atoms not covered by that table. It uses van der Waals radii from Bondi (1964), with the common extensions for elements Bondi did not tabulate (H 1.10 Å after Rowland & Taylor 1996, Ca 2.31 Å after Mantina et al. 2009, Fe 1.80 Å as a conventional approximation), and is intended to keep exploratory calculations usable, not to replace a reviewed parameter set for nonstandard chemistry. FastSASA covers 16 elements (src/fastsasa_radius.c); any other element has no fallback and is rejected.

Element	Radius (Å)	Element	Radius (Å)
H / D	1.10	Br	1.85
C	1.70	I	1.98
N	1.55	Na	2.27
O	1.52	Mg	1.73
F	1.47	K	2.75
P	1.80	Ca	2.31
S	1.80	Fe	1.80
Cl	1.75	Zn	1.39

Glycans And Lipids

ProtOr covers amino acids and nucleic acids. Sugar and lipid atoms fall back to element radii with a warning.

For glycans, FastSASA ships an opt-in extension:

fastsasa --hetatm --config-file share/protor_glycans.config glycoprotein.pdb

protor_glycans.config is the full ProtOr table plus common pyranose residues (GLC, BGC, MAN, BMA, GAL, GLA, NAG, NDG, FUC, FUL, XYS, XYP) mapped onto the existing ProtOr chemical types. It introduces no new radius values, so any atom that was already covered produces exactly the same result as the default table; only previously-guessed sugar atoms change (typically from the element carbon radius 1.70 to the ProtOr aliphatic 1.88). That is the compatibility contract for any future table extension: new entries may only map additional residue/atom pairs onto existing published types, never alter an existing assignment.

For lipids there is deliberately no bundled table: membrane force fields use their own atom naming and their own radii, so the right radii come from the simulation force field, not from a structure-database table. Generate a config from your force field (or, inside VMD, use the VMD integration's -radii vmd mode, which exports the loaded molecule's radii automatically), or rely on element guessing when approximate lipid radii are acceptable.

Relative-SASA (RSA) output uses the bundled ProtOr residue references. They are reported whenever the loaded config declares name: ProtOr, which the bundled share/protor.config does, whether it was auto-discovered, named by FASTSASA_DEFAULT_CONFIG, or passed with --config-file. Any other table, including the glycan extension (name: ProtOr-Glycans), reports N/A for the relative fields, because the references describe ProtOr radii only.

MD Residue Names

Force-field topologies name protonation states, tautomers, and disulfide-bonded cysteine differently from the PDB. FastSASA resolves the common variants to the standard residue for radius lookup, polar/apolar classification, and RSA references, in the CLI, the trajectory command, the Python RadiusConfig, and the feature extraction's per_residue_rsa:

Variant	Standard	Variant	Standard
HID, HIE, HIP	HIS	CYX, CYM	CYS
HSD, HSE, HSP	HIS	ASH	ASP
LYN	LYS	GLH	GLU
ARN	ARG

An exact residue atom entry in the config always wins, so a custom table can still give a variant its own radii. Other nonstandard residues get the element fallback for radii and N/A/NaN for relative SASA; the Python API warns once per unknown residue name.

Reference

J. Tsai, R. Taylor, C. Chothia, and M. Gerstein, “The packing density in proteins: standard radii and volumes,” Journal of Molecular Biology 290, 253–266 (1999), doi:10.1006/jmbi.1999.2829.

A. Bondi, “van der Waals Volumes and Radii,” Journal of Physical Chemistry 68(3), 441–451 (1964), doi:10.1021/j100785a001.

R. S. Rowland and R. Taylor, “Intermolecular Nonbonded Contact Distances in Organic Crystal Structures,” Journal of Physical Chemistry 100(18), 7384–7391 (1996), doi:10.1021/jp953141+.

M. Mantina, A. C. Chamberlin, R. Valero, C. J. Cramer, and D. G. Truhlar, “Consistent van der Waals Radii for the Whole Main Group,” Journal of Physical Chemistry A 113(19), 5806–5812 (2009), doi:10.1021/jp8111556.