libfmftsaxs

Provides tools for manifold-FFT SAXS calculation.

---

Requirements

• libmol2
• gmp
• fftw3
• Check
• Doxygen
• CMake (version > 3.0)

Installation

After cloning the repository, do

$ cd libfmftsaxs
$ mkdir build && cd build
$ cmake -DBUILD_TESTS=True -DBUILD_DOCS=True -DMPI=False -DCMAKE_BUILD_TYPE=Release  ..
$ make
$ make test && make doc
$ make install

If you are planning to use MPI, you can optionally add -DMPI=True. To add verbosity -DVERBOSE=True.

---

Capabilities

This is a library for calculation of SAXS on a single protein molecule as well as on a dimer. It utilizes expansion of scattering amplitudes into spherical harmonics to efficiently compute scattering intensity as described here. The novelty and the main feature of the library, is that it provides functions for SAXS calculation on the heterodimers using FFT (Fast Fourier Transform), representing the dimer's scattering intensity as a correlation function, which allows to score millions of dimer conformations very fast.

---

HOWTO

There are several tools and examples included in the directory examples. Each of the scripts there must be launched from this directory. They do not require you to install the library, just to build. Most scripts generally start from specifing mapping file and parameter file, where the former is needed to create the form-factor table and the latter to add parameters of the atoms. They reside in prms directory.

1. Computing SAXS profile run_single_saxs.sh provides you with an example of SAXS profile calculation. It uses files stored in dimer1, which contains unbound models of Lysozyme and PliG, whose complex is reported in PDB as 4G9S. You can specify c1, c2 and lmax - expansion depth (see documentation for the details). After the script is run, the computed SAXS profile is written into dimer1_saxs_profile and you can plot it using

   $ python plot_curves.py dimer1_saxs_profile

2. Scoring SAXS profiles run_naive_scoring.sh gives you an example of how to score dimer conformations using score_ft_naive binary. The conformations are written in the form of ft-file and respective file with rotation matrices. Each conformation is written, transformed into Euler coordinates, then SAXS profile is created for this conformation and minimized through L-BFGS-B algorithm with respect to the parameters c1, c2, provided the reference (experimental) profile. You can specify the output, where the scores will be written. They are written the form of a table with the following columns: | FT index | SAXS-score | c1 | c2 |.

3. Converting FT-file to Euler angles. run_ft2euler.sh converts ft-file from Cartesian to Euler coordinates and writes it into euler_list in the form: | $z$ | $\beta_{rec}$ | $\gamma_{rec}$ | $\alpha_{lig}$ | $\beta_{lig}$ | $\gamma_{lig}$ |. This means, that receptor is rotated by $(0.0, \beta_{rec}, \gamma_{rec})$ and ligand is rotated by $(\alpha_{lig}, \beta_{lig}, \gamma_{lig})$ and translated along z-axis by $z$.

4. Ultra-fast FFT-SAXS scoring. run_correlate.sh demonstrates the main feature of this library. It uses the protein complex dimer2, which contains unbound receptor and ligand of 1A2K complex. First SAXS curve is calculated for the native conformation of the dimer for a chosen pair of parameters c1, c2 (1.0, 1.0 by default). Then we pretend, that this curve is experimental and feed it to the executable correlate together with the 3 ft-files, which in total contain 210000 conformations. The ft-files are concatenated and fed to correlate as a single list. It uses FFT to score the conformations in super-fast fashion and provides a happy user with the output of the form | Serial number | FT index | SAXS-score | c1 | c2 |, which is then sorted by the serial number (line index in the concatenated list) and split into three lists with SAXS-scores, each of which corresponds to the one of the ft-files. By default the utility correlate is launched with mpirun, but you can use the serial version, if you did not specify MPI flag during the compilation (which will make it proportinally slower).

You can plot_curves.py to build as many curves as you want of a single plot just like this:

$ python plot_curves.py curve1 curve2 ... 

It also computes $\chi$-score between every pair of them.

---

License

See License.TXT for details.

---

Authors

Mikhail Ignatov, Andrey Kazennov and Dima Kozakov.

---

References

1. Svergun, D., C. Barberato, and M. H. J. Koch. "CRYSOL - a Program to Evaluate X-ray Solution Scattering of Biological Macromolecules from Atomic Coordinates." Journal of Applied Crystallography 28.6 (1995): 768-73. Web.
2. Konarev, P. V., M. V. Petoukhov, and D. I. Svergun. "MASSHA - a Graphics System for Rigid-body Modelling of Macromolecular Complexes against Solution Scattering Data." Journal of Applied Crystallography 34.4 (2001): 527-32. Web.
3. Xia, Bing, Artem Mamonov, Seppe Leysen, Karen N. Allen, Sergei V. Strelkov, Ioannis Ch. Paschalidis, Sandor Vajda, and Dima Kozakov. "Accounting for Observed Small Angle X-ray Scattering Profile in the Protein-protein Docking Server Cluspro." Journal of Computational Chemistry 36.20 (2015): 1568-572. Web.