Computational Workflow for Mechanistic Prior Generation: Rutin at the Water-Plant-Virus Interface

Reproducible computational scripts and supplementary outputs supporting a critical mini-review on rutin at the water-plant-virus interface in nature-based wastewater treatment. The integration of computational tools within the four-barrier antiviral phytoremediation model is intended to provide mechanistic priors, identify metabolic pathways, and quantify ligand-binding potentials that are otherwise absent from the literature. This computational layer is designed to guide future empirical LC-MS/MS and infectivity-based confirmation rather than to replace the validation of active species in complex matrices.

Scope: All outputs are hypothesis-generating only and support evidence synthesis for manuscript sections 2-4 and Supplementary Information (S1-S4). They do not predict treatment performance or substitute for infectivity assays in complex environmental matrices.

---

Four-Barrier Antiviral Phytoremediation Model

The review organises evidence and computational outputs around a four-barrier model of virus attenuation in constructed wetland systems:

Barrier	Description	Workflow connection
B1 Sorption and filtration	Virus retention at biofilms, organic matter, roots	Descriptor polarity gradient (Table S2)
B2 Rhizosphere-mediated attenuation	Dissolved rutin / metabolites acting on retained virions	BioTransformer products (Table S3)
B3 Plant internalisation	Virus and rutin co-localisation in tissues	Species pool for future assay design
B4 Intracellular transformation	Metabolism shifts active species	Docking screen: parent vs. products (Table S4)

Computational outputs (descriptors -> metabolites -> docking affinities) are mapped to B1-B4 to identify which species and compartments should be prioritised in bench-scale experiments. They do not predict log-reduction or replace infectivity assays.

---

Repository layout

.
+-- data/
|   +-- species.csv                  # Canonical SMILES source for all ligands
+-- scripts/
|   +-- rutin_insilico_descriptors_core.py   # Step 1: RDKit descriptors
|   +-- rutin_vina_docking_prep.py           # Step 2: Receptor + rutin PDBQT prep
|   +-- rutin_multiligand_vina_1MSC.py       # Step 3: Multi-ligand docking
|   +-- biotransformer_fetch_rutin.py        # Optional: BioTransformer API
|   +-- biotransformer_merge_*.py            # Optional: merge BT outputs
|   +-- README-docking-tools.md             # Vina / Qvina / obabel install notes
+-- tests/
|   +-- test_smoke.py                # pytest: descriptors, SMILES, docking CSV
+-- workflow_outputs/
|   +-- 02_analysis/                 # Generated outputs (not re-committed after release)
+-- Supplementary.md                 # Tables S1-S4
+-- environment.yml                  # Conda environment (recommended)
+-- requirements.txt                 # pip fallback + pytest
+-- Makefile                         # Ordered pipeline (make all / make test)
+-- CITATION.cff                     # Machine-readable citation
+-- CHANGELOG.md
+-- LICENSE

---

Quick start

1 - Install dependencies

Recommended (conda, cross-platform RDKit):

conda env create -f environment.yml
conda activate rutin-review

pip fallback:

pip install -r requirements.txt

RDKit via pip can be unreliable on some platforms. If from rdkit import Chem fails, use the conda route.

2 - Install external tools

Tool	Version	Install
Open Babel	>=3.1.1	conda install -c conda-forge openbabel or download installer
AutoDock Vina	1.2.7	Place binary on PATH or drop vina_1.2.7_win.exe in scripts/

Set OBABEL_EXE / VINA_EXE environment variables if the tools are not on PATH. See scripts/README-docking-tools.md for QuickVina2 and Qvina-W alternatives.

3 - Run the pipeline

# Run all three steps in order
make all

# Or step by step
make descriptors    # Step 1: RDKit descriptors
make docking-prep   # Step 2: 1MSC receptor + rutin ligand
make docking-multi  # Step 3: Multi-ligand Vina vs 1MSC

# Optional: fetch BioTransformer products (requires network + httpx)
make biotransformer

4 - Run tests

make test
# or
pytest tests/test_smoke.py -v

---

Output files and manuscript table mapping

Output	Location	Table
Computational parameters	Supplementary.md Table S1	S1
Molecular descriptors CSV	workflow_outputs/02_analysis/rutin_metabolites_descriptors.csv	S2
BioTransformer products	workflow_outputs/02_analysis/*.json + *.csv	S3
Docking summary	workflow_outputs/02_analysis/docking/rutin_1MSC_multiligand/summary.csv	S4
Run provenance	workflow_outputs/02_analysis/rutin_insilico_manifest.json	S1

---

Reproducibility notes

• All software versions are logged in rutin_insilico_manifest.json at runtime (Python version, RDKit version, UTC timestamp).
• Docking uses exhaustiveness=8 (Vina default) on a blind whole-protomer box; affinity values are not site-matched to Zure et al. 2024 [6,7]. See Supplementary.md Table S1.
• BioTransformer was run in ENVMICRO single-step mode. Cached JSON outputs in workflow_outputs/02_analysis/ are the archival record; re-running the API may return slightly different results as the server is updated.
• SMILES strings are defined once in data/species.csv (canonical source). Scripts embed copies for standalone use; a future refactor should read from this CSV.

---

Citation

Please cite both the manuscript and this repository:

@article{zure2026rutin,
  title   = {Rutin at the Water--Plant--Virus Interface: A Critical Mini-Review
             for Nature-Based Wastewater Treatment},
  author  = {Zure, Diaiti and Rahim, Abdul and Kuo, Hsion-Wen David},
  journal = {Bioresource Technology},
  publisher = {Elsevier},
  year    = {2026},
  note    = {Submitted; currently with editor (March 2026)}
}

See CITATION.cff for the machine-readable form (GitHub will render a "Cite this repository" button automatically).

---

License

MIT - see LICENSE.