logbook/analysis/20250624-als831-macrodomain-analysis/

[giscus[bot]]

logbook/analysis/20250624-als831-macrodomain-analysis/

A logbook for the Diffuse Project

https://diffuse.science/logbook/analysis/20250624-als831-macrodomain-analysis/

[spmeisburger]

Here are the data processing details (H8_8 dataset):

1. Set up directories

ln -s /nfs/chess/scratch/user/spm82/als_20250624/macro/6-23-25/Plate1_H8 data # symlink to directory containing cbfs
mkdir -p dials_360
mkdir -p dials_background
mkdir -p mdx2

2. Run DIALS

cd dials_360
xia2 image=../data/H8_8_00001.cbf:1:3600
cd ..

3. DIALS import background frames

cd dials_background
dials.import ../data/H8_air_10_*.cbf image_range=1,360
cd ..

4. Run mdx2

I used a 64 core machine. If you're repeating this, adjust --nproc as needed.

cd mdx2
mdx2.import_data ../dials_background/imported.expt --chunks 5 211 493 --nproc 5 --outfile bkg_data.nxs
mdx2.bin_image_series bkg_data.nxs 5 20 20 --valid_range 0 100 --outfile binned.nxs --nproc 64
mdx2.import_data ../dials_360/DataFiles/AUTOMATIC_DEFAULT_scaled.expt --chunks 50 211 493 --nproc 5
mdx2.import_geometry ../dials_360/DataFiles/AUTOMATIC_DEFAULT_scaled.expt
mdx2.find_peaks geometry.nxs data.nxs --count_threshold 20 --nproc 64
mdx2.mask_peaks geometry.nxs data.nxs peaks.nxs --sigma_cutoff 3 --nproc 64
mdx2.integrate geometry.nxs data.nxs --mask mask.nxs --subdivide 2 2 4 --nproc 64
mdx2.correct geometry.nxs integrated.nxs --background binned.nxs
mdx2.scale corrected.nxs --mca2020
mdx2.merge corrected.nxs --scale scales.nxs
cd ..

5. Generate visualizations and statistics

This was done using jupyter notebooks. I copied and modified the examples from the insulin-multi-crystal tutorial that comes with mdx2. See for example: https://github.com/diff-use/mdx2/blob/main/examples/insulin-multi-crystal/visualization.ipynb