This repository contains everything needed to reproduce the results of my master's thesis:
Linking data and model quality in X-ray Crystallography
Successfully running the scripts in this repository will require at least 2 GB of free space, as well as much processing power and / or time.
For details on the purpose of this project, please look at master_thesis_adler.pdf
To reproduce my results, make sure to have numpy==1.19.1 and matplotlib==3.3.0 installed, then clone this repository and run new_model.py in the models directory.
Once the new directory exists, move there and enter "sh bg_simulate_all &", then enter "disown", in order to simulate everything in the background; if you do this, you can close the console window while you wait; the simulations will likely take several hours (or days); bg_simulate_all will execute the following scripts: create_ensemble.sh: this will create 20 models and "shake" the coordinates for each one, then combine them into an "ensemble"; create_reference.sh: this will create a reference file to ensure correct indexing by XDS; adjust_max.py: this will determine maximum values for each parameter based on CC1/2; simulate.py: this will simulate frames, then reduce and refine the data with an adjustable number of values (default: 20) for each parameter, and save relevant data statistics in the "datafiles" directory; plot_all.py: this will create plots for the contents of the "datafiles" directory;
You can check progress by looking at the logs in the "logfiles" directory. If you want to skip / change any of the steps, simply alter the relevant lines in bg_simulate_all, or run each script individually.
To simulate data for a specific parameter value, use "fix.py";