Fast SAXS Profile Computation with Debye Formula

Help Topics

Input Fields
Advanced Input Fields
Minimal Ensemble Search

Input Fields

FoXS requires only one mandatory input parameter, the input molecule:
  • Input Molecule: A user has to specify PDB id and chain of the molecule (pdb:chain, e.g. 1ukr:A) or upload a file containing the structure of the molecule in PDB format.

    It is also possible to upload a zip file with several PDB structures. In this case each structure will be fitted to the profile separately. Only zip compression is currently supported. Note that profile calculation is performed for up to 100 PDB files. If you need more extensive profile evaluations please use download version.

    For NMR structures please select suitable MODEL reading option from advanced parameters.

    Please use standard PDB atom naming in the files you upload!

    Hydrogen atoms for proteins and nucleic acids are considered implicitly by adding their form factors to those for their bound heavy atom. If your structure includes other groups, such as lipids, sugars etc., we recommend adding all the hydrogens and turning off "implicit hydrogens" option of FoXS to achieve higher accuracy. Hydrogen atoms can be added with Reduce.

    Please make sure there are no missing residues/nucleic acids in the structure you upload!
    Your structure should be as close as possible to the sample you used to collect the experimental data!
    It is highly recommended to model missing loops, disordered N- and C- termini and His tags.

  • Experimental profile: A user can upload a file with experimental profile (optional). In this case the theoretical profile will be fitted to the experimental one. The experimental file should have 3 columns: q, I(q) and error:

    # q intensity error
    0.00000 3280247.73 1904.037
    0.00060 3280164.59 1417.031
    0.00120 3279915.19 1840.032
    0.00180 3279499.57 1566.084

    Header lines should be preceded by #.

    Please note that: q = (4π sin θ) / λ,
    where 2θ is the scattering angle and λ is the wavelength of the incident X-ray beam.

    Sample input file for lysozyme (PDB code 6lyz) can be found here.
    If your experimental profile does not contain an error column, the server will add 5% error with Poisson distribution.
    Note, lines with zero intensities are skipped. If you profile contains negative intensity values, FoXS will read in the profile until it encounters first negative intensity value.

  • e-mail address: An e-mail address for receiving the link to the results page (optional). The results page will be available for at least a week.

Advanced Input Fields

Setting these parameters is optional:
  • Maximal q value The theoretical profile will be computed up till this q value (default = 0.5)

  • Profile size Number of discrete points that represent the profile (default = 500)

  • Hydration Layer Set hydration layer parameter c2

  • Excluded volume Set excluded volume parameter c1

  • Implicit hydrogens Consider hydrogens implicitly

  • Residue level computation Perform coarse grained profile computation for Ca atoms only. This is not as accurate as full atom calculation, can be used only for large structures

  • Background Adjustment This option applies background adjustment to the input experimental profile. As a result the profile is corrected at higher q values, the intensity values are decreased (see Ciccariello et al. (1988) for details)

  • Offset Use constant offset in profile fitting

  • MODEL reading FoXS supports all possible options of handling PDB files with MODEL record. By default MODEL records are ignored and all ATOM records are read into a single structure. Alternatively it is possible to read in first MODEL only or all the MODELs into separate structures


Your browser does not support the HTML 5 canvas element

PDB file show/hide
# atoms
fit file

The input experimental profile is shown with black dots, while the theoretical profile is a red line. The file with the theoretical profile fitted to the experimental one can be downloaded using the link under the plot.

The quality of the fit, χ value, is printed below the plot.

There are 2 free parameters for profile fitting: (i) c1 - controls the exluded volume of the molecule and (ii) c2 - controls the density of the water layer around the molecule. The default values are c1 = 1.0 and c2 = 0.0. Given experimental profile, FoXS finds a combination of c1 and c2 that achieves the best fit (lowest χ value) to the experimental data.


If multiple PDB files were uploaded by the user, in addition to profile calculation for each structure, the server will also run enumeration and fitting of multiple structures to the input profile. Here is an example

Minimal Ensemble Search

Minimal Emsemble Search (MES) is a previous version of multiple structure fiting.

NEW! Search for the minimal ensemble with the MES (solutions number, from 2 to 5)

The user has to specify subset size, which is the number of conformations out of input PDB files, that will be used for averaging during the search for minimal ensemble. After the minimal ensemble is found, MES output will be displayed:

MES (Minimal Ensemble Search) Fit χ = 2.000
File name Weight
model4m_xx.dat 0.29
model2m_xx.dat 0.39
model3m_xx.dat 0.32
Single PDB Best Fit χ = 5.108
model3m_xx.dat 1.00

The output table displays the filenames of the selected PDB files with conformations. In this example model4m, model2m, model3m were selected with weights of 0.29, 0.39 and 0.32 respectively. The χ value for the fit of the three conformers is 2.0. Next, the table displays the χ value for the best fitting single conformer (model3m, χ = 5.108).
The left plot displays the fit between the profiles. The input experimental profile is shown with red dots, the theoretical profile of the best fitting single conformer is a green line and the averaged profile of the minimal ensemble is a blue line. The right plot displays fit residuals of single conformer compared to the minimal ensemble. Fit residuals is defined by dividing experimental intensities by the modeled intensities. Ideally the values should be close to 1.0. This plot helps to see in which regions of the profile there was an improvement as a result of using an ensemble of conformations instead of single conformation.