Available data
--------------

This folder contains data and scripts to generate a comparison figure of HSP17.1 and 18.1 SAXS signals, and their classification according to different assembly topologies.

The following two Python scripts are available:
- compare_plot.py: generates a superimposition of all experimental signals at different concentrations (in supplemental figure)
- SAXS_classifier.py: classifier of SAXS experimental signals using randomly generated, clash-free assemblies as training examples.

Folders HSP17-7 and HSP18-1 contain data to generate topology classification figure in supplemental information for WT2 and WT1, respectively:

HSP17-7 folder contains:
- conc*dat: SAXS signals
- modloop17_7_cter.pdb: HSP17.7 dimer featuring crystallin domain with C-ter peptide bound
- model_Nter_minimized.pdb: model of N-ter dimer

HSP18-1 folder contains:
- conc*dat: SAXS signals
- 181_dimer.pdb: HSP18.1 dimer featuring crystallin domain with C-ter peptide bound
- 3hhw_Nter_dimer.pdb: model of N-ter dimer

Folder biobox contains the in-house structural biology package biobox, required for execution of the SAXS_classifier.py script.


Requirements
------------

All scripts require Python 2.7.x.
compare_plot.py requires numpy and matplotlib Python packages.
SAXS_classifier.py requires sklearn, numpy, matplotlib Python packages.
Furthermore, the in-house package BiobOx is also required, and provided here (make sure this folder is in your PYTHONPATH).
To setup BiobOx for your computer then enter in biobox folder and type:

python setup.py install

All scripts have been tested on Linux.


Execution
---------

To generate data for HSP17.7, enter in HSP17-7 folder and type:
../SAXS_classifier.py modloop17_7_cter.pdb model_Nter_minimized.pdb conc1.dat saxs

For HSP18-1 data, enter in HSP18-1 and type:
../SAXS_classifier.py 181_dimer.pdb 3hhw_Nter_dimer.pdb conc1.dat saxs

By default, SAXS_classifier.py will process pre-generated data contained in *Nter.dat files within the local folder.
These files provide information about the structure and associated saxs of clash-free symmetrical models.
Data can be generated from scratch, i.e. new assemblies can be indentified and their SAXS collected, by deleting (or moving) *Nter.dat files from their containing folder.
To generate data from scratch, the ATSAS suite should be available in the computer, and the environment variable ATSASPATH set to its main folder.

Forthcoming publications are focused specifically on the our method to classify SAXS signal, and on the BiobOx package.



