3D-Pharmacophore Search
Neural Subgraph Matching 
Shape-Feature Dual Scoring 
Billion-Compound Scale 3D pharmacophore hypotheses are only as valuable as the compounds they retrieve. We execute high-throughput 3D searches across billion-compound libraries with AI-accelerated matching and conformer alignment.
Why 3D-Pharmacophore Search Is the Critical Bridge Between Hypothesis and Hit?
A pharmacophore model defines the binding features, but the database search determines what you synthesize. For seed-stage biotechs, a slow or shallow search misses patent-clear scaffolds. For pharma teams, brute-force docking of billion-compound libraries is computationally prohibitive. 3D-pharmacophore search filters libraries by feature and shape complementarity in minutes, not weeks.
What Sets the Platform Apart
Neural Subgraph Matching
GNN encoders trained on 1.2 million ChEMBL compounds match query pharmacophores to database molecules in embedding space. Orders of magnitude faster than brute-force alignment.
Shape-Feature Dual Scoring
Feature sphere overlap combined with shape complementarity scoring reduces false positives from feature-only matches.
Billion-Compound Scale
GPU-accelerated parallel search across billion-compound libraries. Structure-Based Virtual Screening (SBVS) and Ligand-Based Virtual Screening (LBVS) integration.
The 3D-Pharmacophore Search Suite
Feature-Matching Search
3D Pharmacophore Feature Alignment
- Feature Sphere Matching — Database compounds aligned to query pharmacophore feature spheres (H-bond donors, acceptors, hydrophobic, aromatic) with geometric tolerance.
- Partial Match Scoring — Compounds ranked by number of matched features and spatial deviation, allowing partial matches for exploratory screening.
- Ideal For — Hit Identification enrichment; scaffold-hopping from known actives; Ligand-Based Pharmacophore Generation follow-up.
For virtual biotechs with a validated pharmacophore, feature-matching search retrieves structurally diverse hits from billion-compound libraries in hours. For pharma teams, partial match scoring surfaces analogs that preserve essential features while exploring peripheral substitution space.
Shape-Complementarity Search
Molecular Shape Overlay Screening
- Shape Overlay Scoring — 3D shape similarity between query and database compounds using Gaussian volume overlap and moment-based descriptors.
- Exclusion Volume Enforcement — Forbidden regions defined by the binding pocket or known steric clashes filter out shape-incompatible hits.
- Ideal For — Series rescue when potency is maintained but ADMET fails; Lead Optimization shape constraint enforcement; selectivity profiling against off-target homologs.
Shape matters as much as features. For biotechs facing scaffold fatigue, shape-complementarity search identifies structurally novel compounds that fit the same binding pocket envelope. When combined with ADMET Prediction & Modeling, this delivers developable leads with validated shape compatibility.
Conformer-Aligned Search
Multi-Conformer Pharmacophore Matching
- Conformer Ensemble Sampling — Up to 300 conformers per compound generated and aligned to the query pharmacophore.
- Minimum RMSD Scoring — Best-matching conformer selected per compound, with RMSD and strain energy reported.
- Ideal For — Flexible molecules with multiple bioactive conformations; macrocycle and bridged scaffold screening; Protein-Ligand Docking (Rigid / Flexible / Induced Fit Docking) pre-filtering.
Static single-conformer searches miss flexible hits. Our conformer-aligned search samples low-energy conformations and reports the best pharmacophore fit, ensuring flexible scaffolds are not discarded due to incorrect starting geometry.
Platform Instrumentation
| Software / System | Core Capability |
|---|---|
| PharmacoMatch | Neural subgraph matching for efficient 3D pharmacophore screening via GNN embedding space encoding. |
| LigandScout | 3D pharmacophore database search with feature matching, shape overlay, and conformer alignment. |
| Schrödinger Phase | High-throughput 3D pharmacophore screening with shape constraints and partial feature matching. |
| MOE Pharmacophore | Multi-conformer search with pharmacophore fit scoring and shape complementarity evaluation. |
| Discovery Studio | 3D database search with pharmacophore queries, shape constraints, and exclusion volumes. |
| RDKit + CDPKit | Conformer generation (CONFORGE) and pharmacophore feature computation for custom search pipelines. |
| NVIDIA A100 GPU Cluster | Parallelized billion-compound pharmacophore search and neural embedding inference. |
| PyMOL + Maestro | Hit visualization, pharmacophore alignment inspection, and conformer comparison. |
Standardized Workflow
Project Workflow
A standardized, milestone-driven execution system. From pharmacophore query to ranked hit list—managed by a single computational project team, tracked in real time.
01 Query Review & Database Selection
- Pharmacophore query import from Ligand-Based or Structure-Based models.
- Database selection: Enamine, ChEMBL, ZINC, or custom corporate library.
- Query validation against known actives.
Deliverable: Validated query + database coverage report.
02 Search Mode Configuration
- Search mode selection: feature-matching, shape-complementarity, or conformer-aligned.
- Tolerance sphere optimization and exclusion volume definition.
- Conformer generation parameters (number, energy window, RMSD cutoff).
Deliverable: Search configuration with tolerance and shape parameters.
03 3D Search Execution
- Database search execution across billion-compound libraries.
- Feature matching and shape overlay scoring.
- Partial match filtering and consensus ranking.
Deliverable: Ranked hit list with fit scores and shape metrics.
04 Hit Refinement & Validation
- Hit clustering and scaffold diversity analysis.
- Pharmacophore fit rescoring and Binding Free Energy Calculation (FEP/TI, MM/PBSA) for top hits.
- Molecular Dynamics (MD) Simulations for pose stability (if contracted).
Deliverable: Refined hit dataset with stability metrics.
05 Report & Handoff
- Final search report with ranked hit list.
- Conformer alignment quality and synthetic accessibility.
- Direct handoff to Hit Biophysical Characterization or Co-crystallization and Soaking if contracted.
Deliverable: Final report + data package + transition plan to Hit to Lead or Lead Optimization.
Sample Requirements
| Requirement | Details |
|---|---|
| Pharmacophore query | 3D pharmacophore model in PMZ, PHAR, or SDF format; feature definitions and tolerance spheres |
| Compound database | Enamine, ChEMBL, ZINC, or custom library; specify preferred vendor or in-house collection |
| Known actives | Reference compounds for search validation and enrichment benchmarking |
| Project scope | Hit identification, scaffold-hopping, or lead optimization |
| Prior screening data | Existing virtual screening or docking results to avoid redundancy |
Standard Deliverables
- Validated 3D pharmacophore search query with feature definitions and tolerance spheres
- Ranked hit list (top 100–500) with pharmacophore fit scores and shape complementarity metrics
- Database coverage report and search statistics
- Conformer matching analysis with RMSD and strain energy values
- Electronic data package formatted for Ligand-Based Virtual Screening (LBVS) or Hit to Lead handoff
Frequently Asked Questions
Case Study
Case Study: PharmacoMatch — Neural Subgraph Matching for Billion-Compound 3D Pharmacophore Screening
Published Evidence:
Rose D, et al. PharmacoMatch: Efficient 3D Pharmacophore Screening via Neural Subgraph Matching. arXiv preprint. 2024;arXiv:2409.06316.
Key Findings:
- Neural Subgraph Matching: Reinterprets 3D pharmacophore screening as an approximate subgraph matching problem, encoding query-target relationships in a GNN embedding space.
- Self-Supervised Training: Trained on 1.2 million unlabeled ChEMBL compounds with contrastive learning, requiring no labeled binding data.
- Speed: Vector matching at 0.3 μs per comparison—approximately two orders of magnitude faster than traditional alignment algorithms.
- Performance: Comparable enrichment metrics to existing solutions on virtual screening benchmarks in a zero-shot setting.
Industrial Translation:
For seed-stage biotechs, PharmacoMatch eliminates the computational bottleneck of large-library pharmacophore screening without proprietary training datasets. For pharma teams, the neural embedding approach scales to billion-compound searches while maintaining enrichment quality—directly applicable to Structure-Based Virtual Screening (SBVS) and Ligand-Based Virtual Screening (LBVS) campaigns. Our platform operationalizes this peer-reviewed approach within an audit-ready workflow, pairing PharmacoMatch neural pre-screening with Molecular Dynamics (MD) Simulations and Binding Free Energy Calculation (FEP/TI, MM/PBSA) validation.
Figure 1. Structure-based pharmacophore queries of ten targets of the DUD-E benchmark dataset. (Rose D, et al. 2024)
Reference
- Rose D, et al. PharmacoMatch: Efficient 3D Pharmacophore Screening via Neural Subgraph Matching. arXiv preprint. 2024;arXiv:2409.06316.
Need validated 3D pharmacophore search data to advance your hit identification pipeline? Our team can design a search campaign tailored to your pharmacophore query, compound library, and screening goals. Contact our scientific team today to start your project.
