CATH/TED domain visualization (lean)

Minimal tool to visualize HMM vs TED domain boundaries on AlphaFold structures. Analyses live under analyses/.

Setup

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Optional cache dir for AlphaFold PDB files (defaults to ~/.cache/cath-vis):

export CATH_VIS_CACHE_DIR=/path/to/cache

Usage in notebooks

import pandas as pd
from domain_vis import visualize_domains_and_plot

df = pd.read_csv("results/domain_comparison_results_human.processed.annotated.tsv", sep="\t")
visualize_domains_and_plot(df)

• Color toggle: pLDDT or packing density; if a match column exists, a Match/No Match mode is available.
• Search accepts UniProt accessions or full AlphaFold IDs; accessions normalize to AF-<ACC>-F1-model_v4.

Modules

• domain_vis/core.py: cache-aware AlphaFold fetch, boundary parsing/merging, viewer builder
• domain_vis/visualize.py: unified interactive viewer (exported via domain_vis/__init__.py)

Analyses (optional)

Scripts are kept separate in analyses/ to keep the viewer lean:

python analyses/plot_density.py --input data/domain_comparison_results_human.tsv --output density_human.png
python analyses/scatterplot.py --input ted-evalue.txt --outdir results/plots --ymax 80
python analyses/migrate_upgrade_v2.py
python analyses/migrate_annotate.py

Data inputs

• data/domain_comparison_results_human.tsv
• results/domain_comparison_results_human.processed.tsv
• results/domain_comparison_results_human.processed.annotated.tsv

Notes

• AlphaFold PDBs are cached; delete cache files to refresh.
• Boundary strings like 5-42,60-120_200-240 are parsed and merged.