Exploring the theoretical basis for the double-Wilson distribution

Before notebooks can be run, the environment must be set up. Please run the following commands:

git clone https://github.com/Hekstra-Lab/dw.git
cd dw
conda env create --file wilson.yml
conda activate wilson
pip install -e .

This repository contains a set of IPython notebooks, each standalone and exploring a separate idea around the theoretical basis of the double-Wilson distribution. The notebooks are as follows:

• Notebook 1: Normalizing structure factor amplitudes: A notebook using multiple local scaling methods to prepare normalized structure factor amplitudes from real data, along with an accompanying notebook, Notebook 1A: Normalizing structure factor amplitudes for anomalous data for doing the same with anomalous data.
• Notebook 2a: Illustrating the double-Wilson model using synthetic data: A notebook illustrating the double-Wilson model using synthetic data.
• Notebook 2b: The double-Wilson model with measurement error and resolution-dependence: A notebook exploring ways to incorporate measurement error and resolution dependence into the double-Wilson model.
• Notebook 3: Fitting a double-Wilson model to a pair of datasets: A notebook that fits paired data scaled in Notebook 1: Dataset preparation and local scaling. This is accompanied by Notebook 3A: Fitting a Double-Wilson model to an anomalous dataset.
• Notebook 4: Conventions for parametrizing the Double-Wilson distribution: A small notebook for converting between notational conventions for the Rice and folded Normal probability distributions.
• Notebook 5: A model for bivariate priors: A notebook containing derivations of important probability distributions in the double-Wilson model.
• Notebook 6: Global Scaling for Difference Maps: A notebook exploring a formalism for difference maps under the double-Wilson model.
• Notebook 7: The general case: any number of sets of related structure factors: A notebook constructing a formalism that relates more than three sets of structure factor amplitudes in the multivariate Wilson model.

These notebooks rely on files in the following folders:

• dag_img: a folder containing images of directed acyclic graphs, introduced in [Notebook 7]((7_DAG_PGM_structure_factor_distributions.ipynb).
• double_wilson_data: a folder containing MTZ files containing paired datasets for use in Notebook 1, Notebook 1A, Notebook 2b, Notebook 3, and Notebook 3A.
• dw_tools: a python package containing tools used in all notebooks.
• results_figs: a folder containing figure outputs from most notebooks.