The chordata olfactory receptor database

Wei Han, Siyu Bao, Jin‐Tao Liu, Yiran Wu, Liting Zeng, Tao Zhang, Ningmeng Chen, Kai Yao, Shunguo Fan, Aiping Huang, Yuanyuan Feng, Guiquan Zhang, Ruiyi Zhang, Hongjin Zhu, Hua Tian, Zhi‐Jie Liu, Cao Li-na, Xingxu Huang, Suwen ZhaoProtein 2024 · DOI

TL;DR

Researchers have launched CORD, a comprehensive database for chordate olfactory receptors (ORs), significantly expanding the number of species and ORs covered compared to previous resources. This new database integrates diverse data types, including odorants, OR-odorant pairs, and protein models, offering a powerful platform for advancing olfactory research and understanding chemosensory biology.

What Problem Does This Paper Address?

Olfaction, the sense of smell, is fundamental to many organisms, playing crucial roles in foraging, predator evasion, social communication, and reproduction. The initial step in olfaction involves olfactory receptors (ORs) binding to odorant molecules. In chordates, ORs are encoded by a large family of genes called olfrs. While ORs are primarily known for their role in the nose, they are increasingly found in other tissues and implicated in various diseases.

However, annotating and integrating OR data across the vast diversity of chordate species has been a significant challenge. The rapid evolution of olfr genes, coupled with varying genome sequencing quality and the high proportion of OR pseudogenes, makes accurate annotation difficult. Existing databases, while valuable, often lack comprehensive coverage, systematic integration of data, or user-friendly interfaces for in-depth analysis. This gap hinders researchers' ability to explore the complex world of olfaction across different species and its implications for health and disease.

Methods

The CORD database was built upon the Genome2OR tool, which uses a hidden Markov model (HMM) for rapid and sensitive annotation of olfr genes from genomic sequences. This tool has been updated to simplify the annotation process and has been used to exhaustively annotate over a million olfrs from thousands of chordate genomes.

CORD integrates data from various sources, including:

  • Genomic data: Acquired from the NCBI Assembly database.
  • Odorant and OR-odorant pair data: Collected from seven existing databases (ORDB, OlfactionBase, M2OR, OlfactionDB, ODORactor, InAct, and Glass) and The Good Scents Company (TGSC) and PubChem.
  • Species lineage data: Matched and downloaded using the TaxonKit tool.
  • Olfr names: Obtained from UniProt.

The database architecture employs a front-end and back-end separation for independent development. The user interface is inspired by UniProt, utilizing Vue.js and Element Plus for a responsive design. The back-end uses Nginx as a web server, with NebulaGraph for graph data storage and Elasticsearch for full-text search. Data visualization is powered by AntV G2 and ECharts.

Additional analytical tools integrated into CORD include:

  • BLAST: For homology sequence searches.
  • WebLogo: For visualizing sequence conservation patterns.
  • Profile analysis: For detailed residue distribution analysis.
  • Network explorer: For visualizing complex relationships between ORs.
  • Analysis tools: For mapping OR distribution via histograms and heatmaps.
  • Genome2OR annotation tool: For understanding and utilizing the annotation process.
  • High-precision protein structure prediction: Using OpenFold on functional ORs.
  • OR community detection and network embedding: Using the Louvain method and LargeVis algorithm.

Datasets

CORD integrates data from a vast number of chordate species and their olfactory receptors. The database includes data from 2,776 species, encompassing lancelets, jawless fish, jawed fish, amphibians, reptiles, birds, and mammals. It contains 1,176,818 olfrs, comprising 663,380 functional ORs and 513,438 OR pseudogenes. Additionally, CORD features data on 23,690 odorants and 3,118 OR-odorant pairs.

Key Insights and Ideas

  • Olfactory receptors are more than just for smell: The paper highlights that ORs are not confined to olfactory tissues and are increasingly linked to various disease processes, such as atherosclerosis and diabetes. This broadens the scope of OR research beyond sensory perception.
  • The sheer diversity of olfactory receptors: The number of ORs varies dramatically across species, from over 2,000 in elephants to around 400 in humans. This diversity is a key driver for comparative genomics and understanding evolutionary adaptations.
  • Pseudogenes are a significant challenge and opportunity: A large proportion of OR genes are pseudogenes. CORD's ability to conceptually translate these pseudogenes offers a unique resource for studying their potential roles or evolutionary history.
  • Integrated data is crucial for discovery: By bringing together genomic data, odorant information, protein models, and functional annotations, CORD provides a holistic view that can facilitate novel discoveries in chemosensory science.
  • User-centric design enhances accessibility: The intuitive interface and advanced search capabilities of CORD are designed to make complex OR data accessible to a wider range of researchers, including those new to the field.

Key Results

  • CORD is a comprehensive resource covering 2,776 chordate species and 1,176,818 olfactory receptor genes (663,380 functional, 513,438 pseudogenes).
  • It includes 23,690 odorants and 3,118 OR-odorant pairs, significantly more than previous databases.
  • The database offers high-precision protein models for ORs, generated using OpenFold.
  • CORD has achieved a 64% increase in functional ORs and a nearly 40-fold increase in species coverage compared to ORDB.
  • The platform provides advanced analytical tools, including BLAST, WebLogo, and network visualization, to aid in data exploration.

Limitations and Future Work

  • Limitations: The paper does not explicitly detail limitations of the current CORD version, but implicitly, the accuracy of annotation is still dependent on the quality of the underlying genomic data. The conceptual translation of pseudogenes, while successful for 99%, might still have areas of uncertainty.
  • Future Work: The authors plan to integrate experimentally determined OR structures and build odor-OR complex structures using AlphaFold3. They also aim to continuously update OR data from newly sequenced genomes, implement a more systematic classification system based on DNA phylogeny, and introduce a genome browser feature to analyze olfrs and their neighboring genes in their genomic context.

Summary updated

BibTeX
@inproceedings{han2024chordata,
  author = {Wei Han and Siyu Bao and Jin‐Tao Liu and Yiran Wu and Liting Zeng and Tao Zhang and Ningmeng Chen and Kai Yao and Shunguo Fan and Aiping Huang and Yuanyuan Feng and Guiquan Zhang and Ruiyi Zhang and Hongjin Zhu and Hua Tian and Zhi‐Jie Liu and Cao Li-na and Xingxu Huang and Suwen Zhao},
  title = {{The chordata olfactory receptor database}},
  booktitle = {Protein},
  year = {2024},
  doi = {10.1093/procel/pwae050},
}
Figure