OlfactionBase
OlfactionBase
Peer-reviewed research foundation
PROJECT OVERVIEW
Turning fragmented olfaction data into an explorable research platform.
OlfactionBase is a manually curated scientific database that brings together multiple dimensions of olfaction: odors, odorants and odorless compounds, olfactory receptors, odorant–receptor interactions, and associated proteins. The platform combines structured data retrieval with interactive exploration, including an Olfaction Wheel.
3,985
odorant molecules represented in the current public database.
1,124
odorless compounds represented in the current public database.
106
primary odors, further organized into 572 sub-odors.
30
chemical functional-group classifications used for odorants.
2,067+
human and mouse olfactory receptor records represented across the platform.
2,800+
odorant/pheromone-binding protein records across many species.
THE CHALLENGE
The problem was not merely storing data. It was making relationships discoverable.
Core engineering questions
- How should heterogeneous scientific entities be modeled?
- How can many-to-many biological relationships remain queryable?
- How can chemical information be exposed without overwhelming users?
- How can hierarchy and relationships be visualized interactively?
- How can external scientific identifiers and literature evidence remain connected to records?
Olfaction research spans chemistry, molecular biology, receptor biology and scientific literature. Relevant information is naturally distributed across different resources and represented using different identifiers and structures.
The platform therefore had to support a researcher moving in either direction: odor → chemical → receptor → evidence, or receptor → interacting chemical → odor → evidence.
SOLUTION
An interconnected scientific knowledge platform.
The application was designed around a relational model that treats each scientific entity as a reusable building block and links those entities through explicit relationships.
1. Curated scientific entities
Odors, sub-odors, chemical compounds, receptors and associated proteins are represented as structured entities with domain-specific attributes.
2. Relationship-centric data model
Many-to-many relationships connect odors to compounds and compounds to receptors, while evidence and external identifiers provide provenance and cross-reference paths.
3. Multiple discovery modes
Users can search textual identifiers and chemical attributes, browse categories, or start from a biological entity and follow related records.
4. Visual exploration
The Olfaction Wheel turns hierarchical odor classification and many-to-many relationships into an interactive exploration workflow.
ARCHITECTURE
Relational architecture designed for scientific relationships.
│
├── Search / Filters / Browse
├── Olfaction Wheel / Visualization
└── Entity Detail Pages
│
▼
WEB APPLICATION LAYER
│ Laravel + PHP
│ Server-side data access / business logic
▼
RELATIONAL DATA LAYER
│ MySQL
├── Odors / Sub-odors
├── Chemicals / Functional Groups
├── Olfactory Receptors
├── OR–Odorant Relationships
├── OBP / PBP / Chemosensory Proteins
└── Literature / Cross References
│
▼
EXTERNAL SCIENTIFIC REFERENCES
PubMed · PubChem · UniProt · GenBank · ZINC · other resources
MY CONTRIBUTION
Bridging scientific requirements and production-grade software.
My role focused on translating a complex bioinformatics problem into a usable, maintainable web application and data platform.
Application Architecture
Designed and implemented application structures that connect scientific data models with user-facing workflows.
Database Engineering
Modeled relationships among scientific entities and supported efficient retrieval of interconnected records.
Backend Engineering
Implemented Laravel/PHP application logic, data retrieval and domain-oriented workflows.
Frontend Engineering
Built interfaces for search, navigation, detailed records and scientific data discovery.
Visualization
Worked on interactive data visualization patterns, including the Olfaction Wheel experience.
Scientific Data Integration
Helped bridge chemical, receptor, protein and literature-oriented datasets into a coherent application model.
TECHNOLOGY
The engineering stack
| Layer | Technology / Approach | Purpose |
|---|---|---|
| Backend | PHP · Laravel | Application logic, routing, server-side workflows and data access. |
| Database | MySQL | Structured storage and relationship modeling for scientific entities. |
| Frontend | HTML · CSS · JavaScript · Vue.js | Interactive scientific interfaces and client-side experiences. |
| Visualization | D3.js | Interactive relationship and hierarchy visualization. |
| Development | Git · Composer | Source control and PHP dependency management. |
| Scientific integration | PubMed and external biological/chemical identifiers | Evidence and cross-reference connectivity. |