Epic: Ligands Feature for BioNeighbor

[andytriboletti]

Epic: Ligands Feature for BioNeighbor

Summary

Add a first-class Ligands section to BioNeighbor to enable exploration, comparison, and similarity analysis of ligands used in drugs, assays, and biological systems. Ligands act as the bridge between molecules, targets, and bioactivity, and are essential for understanding mechanisms of action and discovering new therapeutic candidates.

This feature will allow users to view ligands, explore structurally and biologically similar ligands, and understand how ligand features relate to targets, assays, and diseases.

---

Why Ligands Matter

• Ligands are the actual interacting entities in biology
• Drugs often share ligands or ligand scaffolds
• Ligand similarity is more chemically meaningful than drug name similarity
• Medicinal chemistry and bioinformatics workflows are ligand-centric
• Ligands naturally support BioNeighbor’s “neighbor discovery” concept

---

Goals

• Provide a dedicated Ligands tab alongside Drugs, Targets, Diseases, and Assays
• Enable ligand similarity based on:
  • Chemical bonds
  • Functional groups
  • Scaffolds
  • Bioactivity profiles
• Support exploration of:
  • Drug-like ligands
  • DNA/RNA-binding ligands
  • Experimental and assay-tested ligands
• Lay groundwork for future assay simulation and target inference

---

Core Features

1. Ligand Entity View

Each ligand should have a dedicated page showing:

• Chemical structure (SMILES / InChI)
• Molecular formula and weight
• Known targets
• Known assays
• Known associated drugs
• Known disease relevance (if available)

---

2. Ligand Similarity (Neighbor Discovery)

Ligand similarity should support multiple independent dimensions:

A. Bond-Level Similarity

• Shared bond motifs
• Aromatic systems
• Hydrogen bond donors/acceptors
• Charge patterns
• Ring systems

Purpose:

Explain why two ligands bind similarly

---

B. Functional Group / Component Similarity

• Shared chemical components (e.g. indoles, purines, xanthines)
• Reusable pharmacophores
• Known medicinal chemistry motifs

Purpose:

Identify reusable chemical building blocks

---

C. Scaffold Similarity

• Core ring systems
• Murcko-style scaffolds
• Decorated vs undecorated cores

Purpose:

Explore chemical space around a core structure

---

D. Bioactivity Similarity

• Ligands tested in similar assays
• Ligands binding similar targets
• Ligands associated with similar diseases

Purpose:

“Ligands that behave similarly in biology”

---

3. Interaction Classification

Where data exists, ligands should be classified as:

• Agonists
• Antagonists
• Inhibitors
• Modulators

This enables:

• Mechanism-based exploration
• Target-specific ligand filtering

---

4. DNA- and RNA-Binding Ligands

Support identification and flagging of ligands known to interact with:

• DNA
• RNA
• Protein–nucleic acid complexes

These ligands should be:

• Clearly labeled
• Contextualized (e.g. historical use, toxicity risk)
• Included for research and educational purposes

---

Data Sources (Explicit)

Primary Sources

• ChEMBL (database dumps preferred over API)

  • Ligand–target–assay relationships
  • Bioactivity measurements

• PubChem (Compound + BioAssay)

  • Large-scale ligand structures
  • Assay results
  • Broad chemical diversity

Secondary / Curated Sources

• Guide to Pharmacology (IUPHAR)

  • Clean agonist / antagonist / inhibitor classifications
  • High-confidence pharmacology data

• PDB / PDBBind

  • Structural data for ligand–target interactions
  • DNA/RNA-binding ligands

---

MVP Scope

• Import a limited ligand dataset (curated subset)
• Display ligand pages with basic metadata
• Enable at least one similarity method:
  • Structural fingerprint similarity OR
  • Shared target similarity
• Add Ligands tab to navigation

---

Future Enhancements

• Multi-dimensional similarity scoring
• Interactive ligand graphs
• Ligand → assay simulation hooks
• Scaffold hopping exploration
• Toxicity and selectivity heuristics
• User-defined similarity weighting

---

Relationship to Other Features

• Targets: Ligands bind targets
• Assays: Ligands are tested in assays
• Drugs: Drugs contain or derive from ligands
• Diseases: Ligands modulate disease-relevant pathways

Ligands serve as the central connective layer across the platform.

---

Success Criteria

• Users can discover ligand neighbors intuitively
• Ligand similarity produces explainable results
• Data sources are transparent and traceable
• Feature integrates cleanly with existing BioNeighbor entities

---

Notes

• This epic intentionally avoids hard dependencies on unstable APIs
• Emphasis is placed on explainability, not black-box prediction
• Feature is designed to scale from educational to research-grade usage