Critique the following statement. Be sure to include a thought provoking question.

A well-documented pitfall comes from benchmarking on the DUD-E dataset: models often show excellent enrichment but largely because analog and decoy bias make the actives easier to recognize than the decoysnot because the scoring truly captures physics. As Chen et al. demonstrated, performance inflated on DUD-E collapsed when biases were controlled, revealing many false positives/overestimated hits that would likely fail experimentally. Rigid-receptor assumptions and limited water treatment further exacerbate pose plausibility while overstating scores.

How to correct (rescore/validate): Use consensus scoring and physics-based rescoring (e.g., MM-GBSA or alchemical free-energy methods) to penalize poses that only "look good" to one function. Deng et al. showed that adding free-energy calculations after docking improves discrimination between true binders and false positives. Crucially, follow with prospective experimental validation (biochemical/biophysical assays) to verify activity.

Strategies to reduce false positives: (1) Ensemble docking against multiple receptor conformations; (2) explicit/structural waters or water-aware scoring to capture key bridging interactions; (3) re-docking known actives and enrichment tests to sanity-check a workflow; (4) short MD refinements of docked complexes before rescoring to weed out strained or unstable poses.

Implications for interpretation: Warren et al shows docking often fails at affinity rank-ordering, so a great score should be treated as a hypothesis, not proof; over-interpretation risks chasing artifacts. Interpret AC score gaps qualitatively, then stress-test top poses with ensemble docking and verify against orthogonal data.

Data validation/cross-method checks: Benchmark workflow on public sets (DUD-E, PDBbind/CASF) to ensure it retrieves known actives and that improvements persist when biases are minimized. Combine docking + MM/GBSA or FEP + experiment; agreement across these layers substantially raises confidence and filters out docking-driven false positives before wet-lab work.