# When Does Context Help? A Systematic Study of Target-Conditional Molecular Property Prediction Source: [When Does Context Help? A Systematic Study of Target-Conditional Molecular Property Prediction]() Collection index: [llms.txt]() Data updated: 2026-04-10T05:00:13.311Z Authors: Bryan Cheng, Jasper Zhang Published: 2026-04-08T01:19:16.000Z ## Abstract We present the first systematic study of when target context helps molecular property prediction, evaluating context conditioning across 10 diverse protein families, 4 fusion architectures, data regimes spanning 67-9,409 training compounds, and both temporal and random evaluation splits. Using NestDrug, a FiLM-based architecture that conditions molecular representations on target identity, we characterize both success and failure modes with three principal findings. First, fusion architecture dominates: FiLM outperforms concatenation by 24.2 percentage points and additive conditioning by 8.6 pp; how you incorporate context matters more than whether you include it. Second, context enables otherwise impossible predictions: on data-scarce CYP3A4 (67 training compounds), multi-task transfer achieves 0.686 AUC where per-target Random Forest collapses to 0.238. Third, context can systematically hurt: distribution mismatch causes 10.2 pp degradation on BACE1; few-shot adaptation consistently underperforms zero-shot. Beyond methodology, we expose fundamental flaws in standard benchmarking: 1-nearest-neighbor Tanimoto achieves 0.991 AUC on DUD-E without any learning, and 50% of actives leak from training data, rendering absolute performance metrics meaningless. Our temporal split evaluation (train up to 2020, test 2021-2024) achieves stable 0.843 AUC with no degradation, providing the first rigorous evidence that context-conditional molecular representations generalize to future chemical space. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Hit discovery (DMTA replay) on DUD-E](): Hit discovery (DMTA replay); DUD-E; Hit Rate: 87.6. - [Molecular Property Prediction on ChEMBL temporal split 2021 (test)](): Molecular property prediction; ChEMBL temporal split 2021 (test); ROC-AUC: 0.849. - [Molecular Property Prediction on ChEMBL temporal 2022 (test)](): Molecular property prediction; ChEMBL temporal 2022 (test); ROC-AUC: 0.838. - [Molecular Property Prediction on ChEMBL temporal split 2023 (test)](): Molecular property prediction; ChEMBL temporal split 2023 (test); ROC-AUC: 82.3. - [Molecular Property Prediction on ChEMBL temporal split 2024 (test)](): Molecular property prediction; ChEMBL temporal split 2024 (test); ROC-AUC: 84.9. - [Molecular Property Prediction on ChEMBL 2021-2024 temporal split overall (test)](): Molecular property prediction; ChEMBL 2021-2024 temporal split overall (test); ROC-AUC: 0.843. - [Virtual Screening on DUD-E 10 diverse targets](): Virtual Screening; DUD-E 10 diverse targets; Mean AUC: 85. ## Original sources and implementation [Original paper]() [PDF]()