🧬 superdarks: Code files

The superdarks codebase is organized as a four-stage pipeline. Each subfolder below is one stage; click into a stage to browse its scripts and to read the stage-specific notes.

1-query_code/ — per-node TM-align runner and its job-submission wrapper, result consolidation, top-hit re-alignment, and coverage × TM-score hit parsing.
2-post_query_analysis/ — UpSet diagram intersection analysis over filtered and unfiltered hits, plus headless 2D/3D score-contour plots.
3-post_query_informatics/ — UniProt / TrEMBL / InterPro enrichment, 7TM+GPCR keyword filtering, BLASTP network construction, go-forward graph traversal, and Clustal-Omega conservation-to-structure mapping.
4-python_wrappers/ — standalone wrappers used at any stage: a Foldseek GUI front-end and an fpocket subprocess wrapper.

⚠️ Disclaimer

This is research-grade code provided for academic and exploratory purposes only. It is not intended for professional or clinical deployment.

The code was developed and extended over many years to support diverse use cases in computational structural biology.
You may encounter commented sections, experimental blocks, or legacy fragments — these reflect the evolving nature of scientific coding.
Use at your own risk. While we strive for accuracy, this code has not undergone rigorous software engineering review.

📥 Getting Started

Most scripts in superdarks/code/ are driven either by a CONFIG block at the top of the file (edit in place) or by getopt flags for HPC submission via LSF bsub. Before running on a new machine:

Open the script’s header and set cluster, queue, and any absolute paths under /projectnb/isomlab/… or /nethome/… to match your environment.
Confirm required cluster-side binaries are on PATH: TMalign, blastp, clustalo, foldseek, fpocket, pigz, and bsub.
Install Python dependencies listed in each stage’s 3rd_party_requirements.txt.

See the project landing page for the full method overview and citations.