EXPERIMENT 02 / PROMISCUITY
Promiscuous PPAR
Find molecules with strong docking scores across all three PPAR receptors within a shared measurement budget.
- Library
- 259,930 molecules
- Paid budget
- 6,000 measurements / run
- Targets
- PPARA + PPARD + PPARG
- Search seeds
- 0 / 1 / 2
THE EXPERIMENT
What are we looking for?
A candidate is scored by its weakest PPAR interaction, with a QED drug-likeness penalty. It counts as discovered only after all three receptor scores have been measured.
Maximize F = −max(PPARA, PPARD, PPARG) − 10 × (1 − QED)
DOCKSTRING supplies docking scores for PPARA, PPARD, and PPARG. Both methods use the same canonical library and QED values.
FINAL QUALITY
Results at the full budget
Means give equal weight to seeds 0 / 1 / 2. Sample SD describes variation across the 3 recorded runs of each method. Bold marks the higher mean.
| Metric | MolScout | Mobley EI Separate | Difference | Ratio |
|---|---|---|---|---|
| Recall@100 | 98.00%± 1.00 pp SD | 77.67%± 8.74 pp SD | +20.33 pp | 1.26× |
| Recall@250 | 91.20%± 0.40 pp SD | 67.87%± 4.77 pp SD | +23.33 pp | 1.34× |
| Recall@500 | 85.53%± 0.70 pp SD | 58.33%± 2.30 pp SD | +27.20 pp | 1.47× |
| Recall@1000 | 76.60%± 0.70 pp SD | 49.23%± 1.47 pp SD | +27.37 pp | 1.56× |
| AUC@100 | 0.597± 0.042 SD | 0.215± 0.016 SD | +0.382 | 2.77× |
| AUC@250 | 0.526± 0.023 SD | 0.183± 0.012 SD | +0.342 | 2.87× |
| AUC@500 | 0.471± 0.012 SD | 0.155± 0.012 SD | +0.316 | 3.04× |
| AUC@1000 | 0.389± 0.003 SD | 0.123± 0.006 SD | +0.266 | 3.16× |
Mean discovery curves
Compare the methods for the top 100, 250, 500, or 1,000 molecules.
Recall
Percentage of the fixed top K molecules discovered with all required measurements. Higher is better.
Top 100
Top 250
Top 500
Top 1,000
AUC
Average fractional Recall@K over the measurements made so far, from 0 to 1. Higher values mean earlier discovery.
Top 100
Top 250
Top 500
Top 1,000
INDIVIDUAL RESULTS
Individual seed results
Each run uses 6,000 charged measurements. Bold marks the higher value; higher is better for both metrics.
| Top K | MolScout | Mobley EI Separate |
|---|---|---|
| 100 | 97.00% | 85.00% |
| 250 | 91.60% | 70.00% |
| 500 | 85.60% | 58.40% |
| 1000 | 76.60% | 48.70% |
| Top K | MolScout | Mobley EI Separate |
|---|---|---|
| 100 | 99.00% | 80.00% |
| 250 | 91.20% | 71.20% |
| 500 | 86.20% | 60.60% |
| 1000 | 77.30% | 50.90% |
| Top K | MolScout | Mobley EI Separate |
|---|---|---|
| 100 | 98.00% | 68.00% |
| 250 | 90.80% | 62.40% |
| 500 | 84.80% | 56.00% |
| 1000 | 75.90% | 48.10% |
| Top K | MolScout | Mobley EI Separate |
|---|---|---|
| 100 | 0.554 | 0.224 |
| 250 | 0.501 | 0.180 |
| 500 | 0.457 | 0.153 |
| 1000 | 0.388 | 0.123 |
| Top K | MolScout | Mobley EI Separate |
|---|---|---|
| 100 | 0.600 | 0.226 |
| 250 | 0.529 | 0.196 |
| 500 | 0.478 | 0.167 |
| 1000 | 0.392 | 0.129 |
| Top K | MolScout | Mobley EI Separate |
|---|---|---|
| 100 | 0.638 | 0.197 |
| 250 | 0.547 | 0.174 |
| 500 | 0.478 | 0.144 |
| 1000 | 0.388 | 0.118 |
MEASUREMENT EFFICIENCY
Budget to match the reference
Each target level is the reference method’s final mean. The recorded crossing is the first budget that reaches that level. Budget reduction compares MolScout’s crossing with the full measurement budget.
| Metric | Reference final level | MolScout crossing | Reference first crossing | Full budget | Budget reduction |
|---|---|---|---|---|---|
| Recall@100 | 77.67% | 3,451 | 5,983 | 6,000 | 42.5% |
| Recall@250 | 67.87% | 3,456 | 5,996 | 6,000 | 42.4% |
| Recall@500 | 58.33% | 3,211 | 5,996 | 6,000 | 46.5% |
| Recall@1000 | 49.23% | 3,463 | 5,999 | 6,000 | 42.3% |
| AUC@100 | 0.215 | 2,334 | 6,000 | 6,000 | 61.1% |
| AUC@250 | 0.183 | 2,390 | 6,000 | 6,000 | 60.2% |
| AUC@500 | 0.155 | 2,418 | 6,000 | 6,000 | 59.7% |
| AUC@1000 | 0.123 | 2,565 | 6,000 | 6,000 | 57.2% |
HOW WE COUNT
Measurements, recall, and AUC
- One measurement
- One charged molecule–target score. Initialization, repeated calls, and unavailable evaluations keep the charges recorded by the experiment.
- Recall@K
- The fraction of the fixed top-K molecules discovered so far. A molecule must have every score required by the objective. Recall is displayed as a percentage.
- Running AUC@K
- Average fractional recall over the paid budget, using left-step integration. Values range from 0 to 1; a higher value means earlier discovery.
- Mean and SD
- Means give equal weight to seeds 0 / 1 / 2. Sample SD describes variation across the 3 recorded runs of each method.
- Budget percentage
- 100 × charged measurements / canonical library molecules. DUDE-Z uses the per-target budget.
- Budget reduction
- 100 × (1 − MolScout crossing budget / full budget). An unreached threshold is reported as an outcome, without extrapolation.
AUC@K(B) = sum of fractional Recall@K(b), for b = 0 … B−1, divided by B
REFERENCE