STRATA — Self-Terminating Remediation Architecture with Tri-modal Auxotrophy

A sleeping engineered E. coli GMO in a biotech lab detects incoming Cr(VI) ions, awakens and biosenses them, reduces Cr(VI) to benign Cr(III) via NemA enzyme, then triggers the holin–endolysin kill switch and dissolves. A new GMO hatches from the incubation chamber to resume monitoring.

This repository accompanies the manuscript:

In Silico Design and Computational Validation of a Tri-Modular Genetic Circuit for Autonomous Detection and Bioremediation of Hexavalent Chromium in Closed-System Microcosms Zubayer Hasan Shaad — Govt. Tolaram College, Narayanganj, Bangladesh (2026).

The manuscript is the primary scholarly record; the contents of this repository are its supplementary computational artefacts — the simulation code, parameter provenance, and generated figures/CSVs that support every quantitative claim in the paper.

The work is entirely computational. No experimental or wet-lab data were generated; all results are derived from the simulation models in simulations/.

This work is entirely computational. No experimental or wet-lab data were generated; all results are derived from the simulation models in simulations/. The deployment claim is contingent on a direct measurement of the per-bp mutation rate in the deployed strain under the deployed growth conditions, which has not yet been performed. See simulations/parameters.py for the provenance of every number in the manuscript. The structural-evaluation pipeline (NemA*2+) follows the literature-decision-gated reporting protocol in simulations/results/literature_check.md, which demotes Vina scores to simulations/results/nema_docking_results_supplementary.csv and reports only geometric observables in the main text.

All simulation code supporting the manuscript is contained in this repository. No experimental or wet-lab data were generated — all results are derived from computational modeling. Parameter sources (literature-derived vs. assumed/estimated) are documented in simulations/parameters.py, which is the single source of truth for every constant used in the paper.

This repository is archived at Zenodo (DOI badge will be added after first Zenodo release).

All model constants — literature-sourced, assumed, or predicted — are centralized and individually cited/labeled in simulations/parameters.py. Each constant is tagged as one of:

Reviewers: opening parameters.py is the fastest way to verify the provenance of every number in the manuscript body.

The repository is organised around the manuscript. The simulations/ tree mirrors the five computational pillars of the paper; everything else is supporting infrastructure (LaTeX source, citation metadata, validator scripts, dependency pin).

.
├── .python-version               # Pinned interpreter (3.11.x)
├── .gitignore                    # Excludes .venv/, __pycache__/, generated artefacts, etc.
├── LICENSE                       # CC BY 4.0
├── README.md                     # This file
├── requirements.txt              # Pinned Python dependencies
├── CITATION.cff                  # Machine-readable citation metadata (CFF 1.2.0)
├── Journal_Manuscript_2026.tex   # LaTeX source of the manuscript (primary scholarly record)
├── scripts/
│   └── scripts_check_tex.py      # Brace / environment-balance validator for the .tex
└── simulations/
    ├── parameters.py             # Centralized, cited parameter set — read this first
    ├── circuit_ode_model.py      # Pillar 1: ODE systems biology simulation (96 h)
    ├── nemA_mutant_kinetics.py   # Pillar 2: WT vs. hypothesized NemA*²⁺ Michaelis–Menten kinetics
    ├── metabolic_burden_model.py # Pillar 3: Ribosome allocation / GSMM (Scott et al. 2010)
    ├── biosafety_mutation_model.py # Pillar 4: Luria–Delbrück evolutionary containment model
    ├── biosafety_sensitivity.py  # Pillar 4b: Sensitivity sweep over mutation rate and volume
    ├── biosafety_lhs.py          # Pillar 4c: LHS global UQ (N = 10,000; 8 input parameters)
    ├── generate_pdf.py           # Renders the submission-formatted PDF (reportlab + fpdf2)
    ├── generate_circuit_diagrams.py # Renders plasmid module diagrams
    ├── structural_pipeline/      # Pillar 5: NemA*²⁺ structural evaluation (PDB 8BPQ)
    │   ├── __init__.py
    │   ├── utils.py              # Shared heavy-atom / contact-cutoff helpers
    │   ├── fetch_nemA_structure.py  # Downloads 8BPQ.cif and 8BPQ_chainA.pdb
    │   ├── literature_check.py   # Pre-docking PubMed decision gate
    │   ├── analyze_active_site.py  # FMN pocket classification (1st/2nd shell)
    │   ├── prepare_docking.py    # WT + ILE328ALA .pdbqt, chromate .pdbqt
    │   ├── scan_contact_disruption.py # Contact-disruption ranker (Kannan 1999)
    │   ├── geometric_placement.py    # scipy.optimize placement of CrO₄²⁻
    │   └── data/                 # 8BPQ.cif, 8BPQ_chainA.pdb(.pdbqt),
    │                             # 8BPQ_chainA_ILE328ALA.pdb(.pdbqt), chromate.pdb(.pdbqt)
    └── results/                  # Generated figures (PNG) and CSV/JSON supporting the paper
        ├── integrated_96h_simulation.png    # Pillar 1
        ├── nemA_mutant_kinetics.png         # Pillar 2
        ├── metabolic_burden_model.png       # Pillar 3
        ├── biosafety_mutation_model.png     # Pillar 4 — 365-day long-run view
        ├── biosafety_mutation_30day.png     # Pillar 4 — 30-day deployment-window view
        ├── sensitivity_analysis.{png,csv}   # Pillar 4b sweep (figure + numbers)
        ├── biosafety_lhs.{csv,png}          # Pillar 4c — 10,000 LHS samples + 3-panel figure
        ├── nema_active_site.{json,png}      # Pillar 5 — FMN pocket composition + figure
        ├── nema_contact_disruption.{csv,png}# Pillar 5 — second-shell mutation ranking
        ├── nema_docking_overlay.png         # Pillar 5 — visual overlay of the placement
        ├── nema_docking_results.csv         # Pillar 5 — geometric observables (main text)
        ├── nema_docking_results_supplementary.csv # Pillar 5 — Vina note (Cr not in Vina)
        ├── structure_provenance.json        # Pillar 5 — PDB/UniProt ID, SHA-256, method
        ├── literature_check.{json,md}       # Pillar 5 — PubMed decision gate
        └── plasmid_module{1,2,3}.png        # Plasmid module diagrams

Local-only paths (gitignored, not part of the submission): assets/ (ORCID logo, figure SVGs) and vina_binary/ (bundled AutoDock Vina 1.2.5 executable, used by Pillar 5 prep steps). The submitted manuscript and its PDF are both reproducible from the tracked source: Journal_Manuscript_2026.tex → simulations/generate_pdf.py.

scripts/scripts_check_tex.py performs a structural sanity check on Journal_Manuscript_2026.tex (brace balance + LaTeX \\begin/\\end environment balance + citation ↔ bibliography cross-check) without invoking a full LaTeX toolchain. It resolves the manuscript path relative to itself, so it runs from any working directory.

# Activate the virtual environment (Windows)
.venv\Scripts\activate

# (First time only) Install all dependencies
pip install -r requirements.txt

# Re-run every computational pillar to regenerate the figures / tables
python simulations/circuit_ode_model.py
python simulations/nemA_mutant_kinetics.py
python simulations/metabolic_burden_model.py
python simulations/biosafety_mutation_model.py
python simulations/biosafety_sensitivity.py
python simulations/biosafety_lhs.py            # Pillar 4c — 10,000 LHS samples

# Render plasmid module diagrams (used in the manuscript figures)
python simulations/generate_circuit_diagrams.py

# Render the submission-formatted PDF of the manuscript
python simulations/generate_pdf.py

# (Optional) Sanity-check the manuscript LaTeX file
python scripts/scripts_check_tex.py

The structural-evaluation pipeline in simulations/structural_pipeline/ is run end-to-end by invoking its individual modules in order: fetch_nemA_structure.py → literature_check.py → analyze_active_site.py → prepare_docking.py → scan_contact_disruption.py → geometric_placement.py. Each module writes its own output into simulations/structural_pipeline/data/ and simulations/results/.

A separate structural pipeline was run to evaluate the feasibility of the hypothesized NemA*²⁺ variant at the protein level, independent of the ODE / biosafety work in Pillars 1–4. The full source lives in simulations/structural_pipeline/ and runs as a six-stage pipeline:

simulations/results/literature_check.md records a PubMed sanity check that decides how the docking/placement results are reported in the manuscript:

Per the pipeline's decision rule, Vina scores are demoted to the supplementary CSV; only geometric observables appear in the main manuscript text:

ILE328 was selected for the in-silico ILE→ALA mutation in Pillar 2. The score bands are all intermediate (0.3 – 0.6), meaning the contact-disruption analysis does not identify a single dominant hotspot; the engineering decision to focus on ILE328 was made on first-shell spatial proximity to the FMN C4a, not on the contact-disruption score alone.

This pipeline is not a free-energy (ΔΔG) calculation. The contact-disruption score is a qualitative screening ranker (Kannan & Vishveshwara 1999, J Mol Biol 292(2):441–464; verified via PubMed PMID 10493887). The geometric placement is a heuristic local optimizer, not a thermodynamic docking score. The ILE328ALA mutation is an in-silico truncation, not a FoldX/Rosetta-predicted design.

simulations/results/structure_provenance.json records the PDB ID, the chain used, the resolution, the UniProt cross-reference, and SHA-256 hashes of the input files, so the structural pipeline is fully reproducible from the versioned source files in this repository.

All managed via the .venv virtual environment. See requirements.txt for pinned versions.

Core packages: numpy · scipy · matplotlib · fpdf2 · reportlab · Pillow

Python version is pinned in .python-version (3.11.x). The project should run on any 3.11+ interpreter; the pin exists so reviewers and Zenodo archivists get a reproducible environment.

Zubayer Hasan Shaad

· GitHub: @zabdax · ORCID: 0009-0002-2322-8553

The manuscript is the primary scholarly record; please cite it (not just the repository) when referencing STRATA's design, methods, or results. The repository is the artefact bundle that supports the manuscript.

The canonical, machine-readable source is CITATION.cff (CFF 1.2.0); GitHub renders a "Cite this repository" button from it automatically. The BibTeX entries below are kept in sync by hand for environments that don't consume CFF (Overleaf, journal submission portals, etc.).

@article{shaad2026strata_manuscript,
  author  = {Shaad, Zubayer Hasan},
  title   = {In Silico Design and Computational Validation of a
             Tri-Modular Genetic Circuit for Autonomous Detection
             and Bioremediation of Hexavalent Chromium in
             Closed-System Microcosms},
  year    = {2026},
  journal = {(under review)},
  url     = {https://github.com/zabdax/STRATA},
  orcid   = {0009-0002-2322-8553}
}

@software{shaad2026strata_code,
  author    = {Shaad, Zubayer Hasan},
  title     = {{STRATA}---Self-Terminating Remediation Architecture
               with Tri-modal Auxotrophy (supplementary code and data)},
  version   = {1.0.0},
  date      = {2026-07-03},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.21168208},  % (DOI assigned on first Zenodo release)
  url       = {https://github.com/zabdax/STRATA},
  orcid     = {0009-0002-2322-8553}
}

The manuscript, code, and supplementary artefacts in this repository are licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0).

You are free to share (copy and redistribute in any medium or format) and adapt (remix, transform, and build upon) the material for any purpose, even commercially, provided that appropriate credit is given. See the LICENSE file for the full license text.