ecoli-k12-drug-model

E. coli K12 Drug Effect Prediction Model

<p align="center"> <strong>Multi-task GNN predicting antibacterial activity, mechanism of action, and protein targets from compound structure</strong> </p> <p align="center"> <a href="https://huggingface.co/MrMufasi/ecoli-k12-drug-model"><img src="https://img.shields.io/badge/HuggingFace-Model-yellow" alt="HuggingFace"></a> <img src="https://img.shields.io/badge/Python-3.9%2B-blue" alt="Python"> <img src="https://img.shields.io/badge/PyTorch-2.0%2B-red" alt="PyTorch"> <img src="https://img.shields.io/badge/License-MIT-green" alt="License"> </p>

---

What It Does

Given any compound's SMILES string, this model predicts three things simultaneously:

Example Predictions

---

Quick Start

Option 1: One-click launcher (recommended)

Requires Python 3.9+ installed.

git clone https://huggingface.co/MrMufasi/ecoli-k12-drug-model
cd ecoli-k12-drug-model
git lfs pull

Then:

• Mac: double-click START_MAC.command
• Windows: double-click START_WINDOWS.bat

Opens at http://localhost:7860

Option 2: Manual setup

git clone https://huggingface.co/MrMufasi/ecoli-k12-drug-model
cd ecoli-k12-drug-model
git lfs pull

python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install -r requirements-app.txt

ECOLI_PROJECT_ROOT=$(pwd) PYTHONPATH=$(pwd) python app.py

Option 3: Python API

from src.inference.predict import predict_drug_effects

result = predict_drug_effects(
    "OC(=O)C1=CN(C2CC2)c2cc(N3CCNCC3)c(F)cc2C1=O",  # Ciprofloxacin
    device="cpu"
)

print(f"MIC: {result['predicted_mic_uM']:.2f} µM")
print(f"MoA: {result['moa_class']} ({result['moa_confidence']:.0%})")
for t in result['top_targets'][:3]:
    print(f"  Target: {t['gene']} ({t['score']:.3f})")

---

Model Architecture

SMILES Input
    │
    ├── GINEConv GNN (3 layers, 256-dim) ──────────── molecular graph topology
    │     └── 17 atom features + 4 bond features
    │     └── triple pooling: mean + max + attention → 768-dim
    │
    ├── Morgan + RDKit Fingerprints (4096-bit) ────── substructure patterns
    │     └── 3-layer MLP → 256-dim
    │
    ├── Physicochemical Descriptors (12 features) ─── scaffold-independent properties
    │     └── MW, logP, TPSA, HBD, HBA, RotBonds, ArRings, Rings, FrCSP3, MolMR, etc.
    │     └── 2-layer MLP → 256-dim
    │
    └── Data Source Embedding (8 → 16-dim) ────────── corrects for inter-source bias
          │
          ▼
    Fusion: 768-dim MLP + LayerNorm + residual connection
          │
          ▼
    ┌─────────────────────────────────────────────────────────────┐
    │ Target Predictor                                            │
    │   Bilinear attention: compound emb × ESM-2 protein embs    │
    │   17 E. coli genes with pre-computed ESM-2 (650M) vectors  │
    └──────────────────────────┬──────────────────────────────────┘
                               │
                               ▼
    ┌─────────────────────────────────────────────────────────────┐
    │ Metabolic Integration Layer (differentiable FBA surrogate)  │
    │   iML1515: 1,516 genes × 2,712 reactions × 40 pathways     │
    │   Computes: reaction inhibition → flux perturbation →       │
    │   pathway impact → growth proxy → 256-dim features          │
    └──────────────────────────┬──────────────────────────────────┘
                               │
                 ┌─────────────┼─────────────┐
                 ▼             ▼             ▼
            MIC Head      MoA Head     Pathway Head
          3-layer MLP    6-class CE    40-dim BCE
          → µg/mL        → class       → pathway impacts

Total parameters: ~9M (7.6M trainable)

---

Training

Curriculum Learning

Phase A builds strong MIC features. Phase B trains the target head without destabilising the GNN. Phase C brings MoA online with differential learning rates.

Loss Function

Regularisation

Test Set Results (v21)

Evaluated on a held-out scaffold-split test set of 2,743 compounds.

---

Dataset

27,488 Compounds from 7 Sources

Data Quality

• Enzyme IC50 exclusion: 2,585 enzyme-level IC50/EC50 records (beta-lactamase, endonuclease assays) removed — these are not whole-cell MIC values
• Variance filter: Compounds with IQR > 2 log₂ units across measurements dropped (203 high-variance compounds)
• Winsorisation: MIC values clipped at 1st/99th percentiles
• Deduplication: Priority order — ChEMBL > SPARK > CO-ADD > PubChem > PATRIC

Labels

MoA Classes

MoA labels expanded from 482 curated → 5,918 via Tanimoto similarity propagation (threshold 0.40).

Predicted Protein Targets

17 targets total (filtered from 1,516 iML1515 genes to those with ≥20 labelled compounds).

Splitting

• Scaffold-based — Murcko scaffold clustering ensures val/test contain novel chemotypes
• MoA-stratified — MoA-labelled compounds distributed proportionally across all splits
• 80/10/10 train/validation/test

---

Biological Integration

ESM-2 Protein Embeddings

Pre-computed embeddings from facebook/esm2_t33_650M_UR50D for all 1,516 iML1515 gene products. Projected from 1280-dim → 512-dim for bilinear attention scoring against compound embeddings.

iML1515 Metabolic Model

The differentiable FBA surrogate uses matrices from the iML1515 genome-scale metabolic model:

• Gene-reaction matrix: 1,516 genes × 2,712 reactions (binary)
• Reaction-pathway matrix: 2,712 reactions × 40 pathways (binary)
• Wildtype FBA fluxes: Baseline metabolic state from constraint-based modelling
• Biomass coefficients: Growth contribution of each reaction
• Learnable sensitivity weights: How much each reaction responds to drug-induced inhibition

---

File Structure

ecoli-k12-drug-model/
├── app.py                          # Gradio web interface
├── train.py                        # Training script
├── install_and_run.py              # One-click installer + launcher
├── START_MAC.command               # Mac double-click launcher
├── START_WINDOWS.bat               # Windows double-click launcher
├── requirements-app.txt            # Inference-only dependencies
├── requirements.txt                # Full training dependencies
├── configs/
│   └── model_config.yaml           # Architecture hyperparameters
├── models/
│   ├── best_model.pt               # Best Phase B checkpoint (val_mic=1.09)
│   ├── best_model_phase_c.pt       # Best Phase C checkpoint (has MoA)
│   └── final_model.pt              # EMA model at end of training
├── src/
│   ├── data/
│   │   ├── merge_datasets.py       # Multi-source dataset builder
│   │   ├── dataset.py              # PyTorch Dataset + collate
│   │   ├── precompute_features.py  # Graph/FP/physchem caching
│   │   ├── compile_moa_labels.py   # 180 curated antibiotic MoA labels
│   │   ├── fetch_chembl_moa.py     # ChEMBL mechanism API
│   │   ├── fetch_patric_amr.py     # PATRIC clinical MIC data
│   │   ├── fetch_bindingdb_targets.py  # BindingDB interactions
│   │   └── propagate_moa_labels.py # Tanimoto MoA propagation
│   ├── features/
│   │   └── molecular_graph.py      # SMILES → graph + fingerprint + physchem
│   ├── models/
│   │   ├── compound_encoder.py     # Multi-branch encoder (GNN+FP+Physchem)
│   │   ├── target_predictor.py     # Bilinear compound-protein scoring
│   │   ├── metabolic_layer.py      # Differentiable FBA surrogate
│   │   └── multitask_head.py       # Top-level model with task heads
│   ├── training/
│   │   ├── losses.py               # Multi-task loss function
│   │   └── evaluation.py           # Test set metrics
│   └── inference/
│       └── predict.py              # Single-compound prediction API
└── data/
    ├── raw/                        # Source data (.parquet files)
    └── processed/                  # Feature caches, splits, gene lists

---

Interpreting Results

MIC Values

The model predicts in log₂ µM internally. To convert:

• µM → µg/mL: multiply by molecular weight / 1000
• log₂ µM → µM: compute 2^(predicted value)

MoA Confidence

The model outputs softmax probabilities across 6 classes. Predictions with >80% confidence on the top class are generally reliable. Lower confidence suggests the compound may have a novel or mixed mechanism.

Target Scores

Target interaction probabilities range 0–1. Most compounds show low activation (<0.1) across all targets. Scores above 0.3 indicate meaningful predicted binding. The model is most confident for well-characterised targets like folA (dihydrofolate reductase) and lpxC (LPS biosynthesis).

---

Limitations

1. Prediction accuracy: MIC predictions have ~2.5-fold average error (RMSE 2.21 log₂ µM). This is comparable to inter-laboratory MIC variability but should be treated as order-of-magnitude estimates, not precise values.

2. E. coli K12 only: Trained on laboratory K12 strain data. Clinical isolates with acquired resistance (ESBL, carbapenemases, plasmid-mediated resistance) will have different MIC profiles.

3. Scaffold bias: Best performance on compounds structurally similar to the training set (fluoroquinolones, beta-lactams, aminoglycosides, tetracyclines, sulfonamides). Novel scaffolds may have higher prediction error.

4. Target coverage: Only 17 of ~1,500 E. coli proteins have sufficient training labels. The model cannot predict interactions with uncharacterised targets.

5. MoA propagation: 91% of MoA labels are inferred by Tanimoto similarity (threshold 0.40), not experimentally confirmed. Novel chemotypes may be misclassified.

6. Not for clinical use. This is a research tool for prioritising compounds in early-stage drug discovery. All predictions require experimental validation.

---

System Requirements

---

Citation

@misc{ecoli-k12-drug-model,
  title={E. coli K12 Drug Effect Prediction Model},
  author={Alex Sheridan},
  year={2026},
  url={https://huggingface.co/MrMufasi/ecoli-k12-drug-model}
}

---

<p align="center"> Built with PyTorch, PyTorch Geometric, ESM-2, RDKit, and Gradio </p>