SASHA
Manual

 

 

General Information

For general information see Reference  The users are referred to this paper for details.

Introduction

The instructions for use of the beta-version of SASHA with an example of a dialogue session commented at each step are given below.

SASHA is a program that allows to determine the low-resolution shape of a homogeneous particle from solution scattering data. No extra information is required.

The program runs in a command line dialogue mode and at present does not use any graphics, so that it is portable to any hardware platform. To plot the results, the User should use other programs e.g. DSPLOT to inspect the quality of fit to experimental data and ASSA (SGI or SUM) or PH(IBM-PC) to draw particle shapes.

SASHA normally works in the USER mode with default answers. The program may stop at the very beginning if it is unable to compute the radius of gyration (Rg) using the Guinier approximation. This is deliberated: if Rg cannot be computed, one should not try shape determination; check your data if this happens.

For typical experimental data, the default (L=4) resolution can be taken. If you apply symmetry restrictions reducing the number of free model parameters, higher harmonics may be used, for example, L=6 with the symmetry Pmmm.

The output fit for monomer data is normally placed onto .fim and .fid for dimer data; the resulting shape Flm coefficients onto .flm. This file contains also detailed information about search conditions and shape parameters, such as integral parameters (radius of gyration, volume), distance between monomer centers, etc.

If your structure is a dimer, you could e.g. treat it as a single body with P2 symmetry. In this case you enter the experimental data as MONOMER curve, and the outputs will be as usual, *.fim and *.flm.

Another option is to look for the shape of the monomer and for the distance between the monomers. Enter the curve as DIMER curve (answer just to MONOMER). The outputs will be *.flm (monomer shape; the distance is written at the bottom of the file; see comments to ASSA how to display a dimer), and *.fid (fit to the data).

If you have the scattering curves of the dimer and of the monomer enter both of them; the program will try to simultaneously fit them. The output files will then be *.flm, *.fim, *.fid.

The fits can be viewed by typing dsplot .fim (or fid), the shapes with ASSA (see corresponding comments to these programs).

Dialogue Session

Each step of the dialogue is commented by lines started with a pair of exclamation marks.

Welcome to SASHA v2.04b from 15.06.99 13:50

Working mode [ User/Developer ] ........ < User >:

At this step, the USER mode is selected by pressing carriage return to give the default answer. Default values are suggested in angular brackets in the prompt line.

MONOMER SAS data file name ............. < .dat >: t30m

Here the file t30m.dat contains experimental scattering intensities from a solution of monomers. See later for the format of this file.

> Number of data points read ........................... : 91
Angular units used in MONOMER data:

4*Pi*SIN(theta)/lambda [1/Ångstrom]	(1)
4*Pi*SIN(theta)/lambda [1/nM]	(2)
2*SIN(theta)/lambda [1/Ångstrom]	(3)
2*SIN(theta)/lambda [1/nM]	(4) .......... < 1 >:

The default answer "1" specifies the scale of angular grid.

DIMER SAS data file name ............... < .dat >: t30d

In this case, there is one more data set measured from a solution of dimers. If only monomer data is available the User should simply press carriage return.

> Number of data points read ........................... : 91
Angular units used in MONOMER data:

4*Pi*SIN(theta)/lambda [1/Ångstrom]	(1)
4*Pi*SIN(theta)/lambda [1/nM]	(2)
2*SIN(theta)/lambda [1/Ångstrom]	(3)
2*SIN(theta)/lambda [1/nM]	(4) .......... < 1 >:

See above.

OUTPUT file name ....................... < .flm >: out

This answer specifies the name of output files.

> MONOMER case:

Guinier fit using 11 points:
Radius of gyration ..................................... : 12.71
I(0) estimate .......................................... : 2.336e+8
Estimated Porod volume ................................. : 1.305e+4

> DIMER case:

Guinier fit using 5 points:
Radius of gyration ..................................... : 17.43
I(0) estimate .......................................... : 9.318e+8
Estimated Porod volume ................................. : 2.773e+4

This is information calculated from experimental data. At this stage, the User may analyse the quality of the data if he knows estimates of the radii of gyrations and volumes of the molecules. The program may stop at this stage if it is unable to compute the Rg using the Guinier approximation. It is deliberated: if Rg cannot be computed, one should not try shape determination; check your data if this happens.

Smax for the MONOMER fitting region .... < 0.5000 >: .35

The default is the maximum value of the scattering vector as specified in the data file. Here, the the range of fitting is reduced to omit the outer portion of the curve.

> Width of the angular region in Shannon channels ...... : 7.640
62 data points in [ 5.000E-02 : .355 ]

the number of Shannon channels is a measure of how many structural parameters describe the shape. The User will use this information later when answering a prompt for the maximum harmonic order, L. The maximum number of model parameters is (L+1)² and decreases for symmetric shapes.
In practice the number of Shannon channels may be 1.5 times less than the number of parameters in the shape model.
If both monomer and dimer curves are handled, the allowed number of parameters increases by a factor of 1.5 (rule of thumb).

Smax for the DIMER fitting region ...... < 0.5000 >: .30

The same choice was made for the dimer data.

> Width of the angular region in Shannon channels ...... : 8.160
52 data points in [ 5.000E-02 : .305 ]

See above.

SEARCH FOR THE BEST ELLIPSOID PARTICLE, MONOMER CASE.

> Half-axes of the best ellipsoid: 18.52 18.52 11.01

SEARCH FOR THE BEST ELLIPSOID PARTICLE, DIMER CASE.

> Half-axes of the best ellipsoid: 25.99 25.99 11.37

At this point, the program has found ellipsoidal shapes that fit the experimental data. Due to the singularity of this problem, two of three axes may coincide but nevertheless these bodies reflect the low-resolution anysometry of the particle. The next messages compare two independent estimates of the particle volumes. The better the data, the closer the estimates.

> MONOMER Porod volume ................................. : 1.305e+4
> MONOMER ellipsoid volume ............................. : 1.581e+4

> DIMER Porod volume ................................... : 2.773e+4
> DIMER ellipsoid volume ............................... : 3.216e+4

Enter initial approximation file or press CR to choose
a default starting particle shape ...... < .flm >:

Here the User may specify the name of a file containing Flm coefficients of the starting model. Pressing CR denotes that the starting approximation will be generated by the program.

Symmetry [P1/P2/P222/Pmmm/P3] .......... < P1 >: p3

This answer specifies the desired particle symmetry. The higher the symmetry, the smaller the number of parameters in the model.

Max. harmonic order for the shape ...... < 4 >: 3

This is an important choice (see comments above). Usually, at the above number of Shannon channels and P1 symmetry one may set L=4, and up to 6 with the Pmmm symmetry.

Specify the shape anysometry:
S - spherical particle;
O - oblate;
P - prolate ......................... < O >: p

If the program generates the starting approximation itself, the user may enter the desired anisometry. Sometimes is is useful to analyse the dependence of the solution on the starting approximation and thus validate the final structure.

WIGRID --- 1627 3j coefficients used

Distance between dimer units (angstroms) < 25.16 >:

 

The default is the value calculated from the volume of the best ellipsoid fitting monomer data.

Adjust the dimer distance [ Y / N ] .... < Yes >:

The default is to adjust the distance during the search.

Monomer fraction in dimer sample [0:1] . < 0.0 >:

Sometimes, due to dissociation, the dimer sample may contain a significant part of monomer. The User may enter here the approximate value and let it vary during the search by answering "YES" to the next prompt:

Adjust the monomer fraction [ Y / N ] .. < No >:

In this example, the User decided not to search the volume fraction.

The next two questions prompt for parameters required for shape smoothness constraints.

Enter the molecular weight in kDa .....................: 30
Enter the monomer Rg in angstroms .....................: 12.5

One may not have this information about the molecule, but do not hesitate to enter rough approximations. Too wrong values may lead to inadequate penalties on the shape features thus resulting in artefacts.

Now the dialogue is finished and we have to wait for the results. The results will written to a set of files named OUT.*.

OK

The output files entitled OUT.* contain:

out.em1 : Fit by monomer ellipsoid
out.ed1 : Fit by dimer ellipsoid
out.fim : Fit by monomer shape
out.fid : Fit by dimer particle
out.flm : Restored shape, Flm coefficients.

The fits are plain ASCII files and can be visualized by any plotting software

To look at the shape on SGI, SUN, DEC: invoke the ASSA system.

To look at the shape on IBM-PC : invoke the MASSHA program

flm2sld out.flm
ph16 out.sld (see brief instructions to ph)

Reference

Svergun, D.I., Volkov, V.V., Kozin, M.B. and Stuhrmann, H.B. (1996). New developments in direct shape determination from small-angle scattering. 2. Uniqueness. Acta Crystallogr. A52, 419-426.