Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
SubDyve: Subgraph-Driven Dynamic Propagation for Virtual Screening Enhancement Controlling False Positive
Jungseob Yi, Seoyoung Choi, Sun Kim, Sangseon Lee
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add LFDR seed refinement (ablation) | SubDyve with subgraph network · CDK7 PU dataset | BEDROC (%): 78.68 -> 83.44 | Improved |
| add LFDR seed refinement (ablation) | SubDyve with subgraph network · CDK7 PU dataset | EF 1%: 89.68 -> 97.59 | Improved |
| add subgraph fingerprint network (ablation) | SubDyve with LFDR · CDK7 PU dataset | BEDROC (%): 63.78 -> 83.44 | Improved |
| add subgraph fingerprint network (ablation) | SubDyve with LFDR · CDK7 PU dataset | EF 1%: 67.22 -> 97.59 | Improved |
| add subgraph fingerprint network without LFDR (ablation) | rdkit + NP · CDK7 PU dataset | BEDROC (%): 79.04 -> 78.68 | No clear effect |
| add LFDR refinement to subgraph network (50 seeds) (ablation) | Subgraph + NP · CDK7 PU dataset (50 seeds) | BEDROC (%): 46.33 -> 51.78 | Improved |
| add LFDR refinement to subgraph network (150 seeds) (ablation) | Subgraph + NP · CDK7 PU dataset (150 seeds) | BEDROC (%): 55.08 -> 59.07 | Improved |
| add LFDR refinement to subgraph network (250 seeds) (ablation) | Subgraph + NP · CDK7 PU dataset (250 seeds) | BEDROC (%): 61.96 -> 66.73 | Improved |
| increase seeds from 50 to 250 (ablation) | SubDyve · CDK7 PU dataset | BEDROC (%): 51.78 -> 66.73 | Improved |
| use SubDyve instead of the PU learning method GRAB | GRAB · CDK7 PU dataset | EF 1%: 45.21 -> 97.59 | Improved |
| use SubDyve instead of PSICHIC | PSICHIC · CDK7 PU dataset | BEDROC (%): 9.37 -> 83.44 | Improved |
| use SubDyve instead of the best fingerprint baseline at 250 seeds (avalon + NP) at EF 0.5% | avalon + NP · CDK7 PU dataset (250 seeds) | EF 0.5%: 86.96 -> 85.44 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DrugCLIP — DrugCLIP: Contrastive Protein-Molecule Representation Learning for Virtual Screening (NeurIPS 2023) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- AANet: Virtual Screening under Structural Uncertainty via Alignment and Aggregation (NeurIPS 2025 · similarity 0.84)
- DMol: A Highly Efficient and Chemical Motif-Preserving Molecule Generation Platform (NeurIPS 2025 · similarity 0.84)
- PharmacoMatch: Efficient 3D Pharmacophore Screening via Neural Subgraph Matching (ICLR 2025 · similarity 0.84)
- M²N: A Progressive Macro-to-Micro 3D Modeling Scheme for Unveiling Drug-Target Affinity (AAAI 2025 · similarity 0.83)
- Monroe: A Molecular Foundation Model for In-context Probabilistic Inference (NeurIPS 2026 · similarity 0.83)
- UniMatch: Universal Matching from Atom to Task for Few-Shot Drug Discovery (ICLR 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
