toxoMine: an integrated omics data warehouse forToxoplasma gondiisystems biology research

David B. Rhee, Matthew McKnight Croken, Kevin Shieh, Julie Sullivan, Gos Micklem, Kami Kim, Aaron GoldenDatabase 2015 · DOI

TL;DR

This paper introduces toxoMine, a new integrated omics data warehouse specifically designed for Toxoplasma gondii research. It addresses the challenge of managing and analyzing the vast, diverse datasets generated by high-throughput biological assays, enabling researchers to explore complex relationships between genetic and epigenetic factors and parasite pathogenicity.

What Problem Does This Paper Address?

Toxoplasma gondii is an obligate intracellular parasite responsible for toxoplasmosis, a significant human health concern. Understanding its virulence and stage transition processes requires integrating data from various high-throughput techniques like epigenomics, genomics, and proteomics. However, the sheer volume and heterogeneity of this data pose a major challenge for collection, organization, and analysis. Researchers need a robust system to effectively query and explore these complex datasets to uncover the underlying genetic and epigenetic mechanisms driving T. gondii's pathogenicity.

Methods

toxoMine is built upon the InterMine framework, an open-source data warehouse solution known for its flexible data model and powerful query engine. The core methodology involves:

  • Data Integration: Integrating diverse datasets, including public T. gondii data (from resources like ToxoDB) and high-throughput experimental data generated by the research group.
  • Ontology-Based Data Modeling: Utilizing the Sequence Ontology (SO) framework to create a structured and searchable data space, allowing for the integration of various biological data types (genomic, epigenomic, proteomic, pathway, etc.).
  • Custom Data Models: Extending the core InterMine data models to accommodate specific high-throughput experiment data types (e.g., project, experiment, submission, microarray, toxoplasma mutant, antibody).
  • Local Staging Database: Employing a local PostgreSQL database as a staging area to organize and curate experimental data and metadata before integration into toxoMine.
  • User-Friendly Interface: Providing a web interface with features like browsing, keyword search, pre-defined "templates" (queries), a "QueryBuilder" for custom query creation, a "Regions" tool for genomic feature analysis, and a "Lists" tool for managing and analyzing gene sets.
  • Interoperability: Leveraging InterMine's InterMOD infrastructure to enable cross-species queries with other InterMine-based systems.
  • Web Services: Offering direct database access via HTTP web services for programmatic querying.
  • Data Upload: Implementing a system for uploading new high-throughput experimental data using a modified BIR-TAB format.

Datasets

toxoMine integrates:

  • Public T. gondii Data:
    • Genome sequences and annotation data (from ToxoDB, release 11.0 for ME49, GT1, and VEG strains).
  • Transcription start site (TSS) region data (from Yamagishi et al.).
  • Protein and protein domain data (from UniProt and InterPro).
  • Homolog data (from OrthoMCL).
  • Gene Ontology (GO) data (from the GO consortium).
  • Publication data (from PubMed).
  • High-Throughput Experimental Data:
    • ChIP-chip data (e.g., identifying genome-wide binding sites for GCN5b).
  • ChIP-seq data.
  • RNA-seq data.
  • Protein-protein interaction data (e.g., identifying GCN5b protein-protein interaction partners).
  • Toxoplasma mutant data.
  • Antibody data.

The paper does not introduce new benchmark datasets but rather focuses on integrating existing and newly generated experimental data.

Key Insights and Ideas

  • The Power of Integration: The paper strongly emphasizes that the true value of toxoMine lies in its ability to integrate disparate omics datasets. This integration allows researchers to move beyond analyzing individual datasets in isolation and instead explore complex, multi-layered biological questions.
  • Democratizing Complex Analysis: By providing user-friendly interfaces like "templates" and "QueryBuilder," toxoMine aims to make sophisticated data analysis accessible to researchers who may not have extensive bioinformatics expertise. This lowers the barrier to entry for complex biological inquiry.
  • Complementing Existing Resources: toxoMine is positioned not as a replacement for, but as a complement to, existing community databases like ToxoDB. It fills a niche by providing a data warehouse specifically for integrative analysis of high-throughput experimental data, enhancing the utility of existing resources.
  • Interoperability as a Force Multiplier: The ability to query other InterMine instances (e.g., through InterMOD) significantly expands the scope of research possible. This allows for cross-species comparisons and leveraging knowledge from better-annotated model organisms.

Key Results

  • Functional Data Warehouse: toxoMine successfully integrates a substantial amount of public T. gondii data with high-throughput experimental data, creating a comprehensive resource.
  • Sophisticated Querying Capabilities: The system allows for complex queries involving multiple datasets, enabling researchers to ask nuanced questions about gene regulation and pathogenicity.
  • User-Friendly Exploration Tools: The paper demonstrates the utility of browsing, keyword search, templates, and the QueryBuilder through examples, showcasing their effectiveness in data exploration.
  • Demonstrated Use Case: The paper provides a detailed demonstration of how toxoMine can be used to analyze ChIP-chip binding site data, illustrating the process of data loading, region analysis, GO term enrichment, and protein domain analysis, leading to the identification of relevant biological insights.

Limitations and Future Work

  • Not the Sole "Portal": The authors acknowledge that toxoMine is not intended to be the sole authoritative portal for all T. gondii omics data, but rather a complementary resource that enhances existing ones.
  • Continuous Data Addition: The team plans to continuously add new datasets and features based on user feedback.
  • RNA-seq and Proteomics Integration: Future work includes incorporating RNA-seq expression data with appropriate analysis widgets and integrating proteomics data, drawing on the experience from MitoMiner.
  • Host-Parasite Interactions: There is a plan to add related host data (human and mouse) to facilitate research on host-parasite interactions.
  • InterMine Variant Interoperability: The team aims to add functionality for toxoMine to interact with InterMine variants beyond those using MOD infrastructure, such as TargetMine, to support drug discovery research.

Summary updated

Abstract

Toxoplasma gondii (T. gondii) is an obligate intracellular parasite that must monitor for changes in the host environment and respond accordingly; however, it is still not fully known which genetic or epigenetic factors are involved in regulating virulence traits of T. gondii. There are on-going efforts to elucidate the mechanisms regulating the stage transition process via the application of high-throughput epigenomics, genomics and proteomics techniques. Given the range of experimental conditions and the typical yield from such high-throughput techniques, a new challenge arises: how to effectively collect, organize and disseminate the generated data for subsequent data analysis. Here, we describe toxoMine, which provides a powerful interface to support sophisticated integrative exploration of high-throughput experimental data and metadata, providing researchers with a more tractable means toward understanding how genetic and/or epigenetic factors play a coordinated role in determining pathogenicity of T. gondii. As a data warehouse, toxoMine allows integration of high-throughput data sets with public T. gondii data. toxoMine is also able to execute complex queries involving multiple data sets with straightforward user interaction. Furthermore, toxoMine allows users to define their own parameters during the search process that gives users near-limitless search and query capabilities. The interoperability feature also allows users to query and examine data available in other InterMine systems, which would effectively augment the search scope beyond what is available to toxoMine. toxoMine complements the major community database ToxoDB by providing a data warehouse that enables more extensive integrative studies for T. gondii. Given all these factors, we believe it will become an indispensable resource to the greater infectious disease research community.
BibTeX
@inproceedings{rhee2015toxomine,
  author = {David B. Rhee and Matthew McKnight Croken and Kevin Shieh and Julie Sullivan and Gos Micklem and Kami Kim and Aaron Golden},
  title = {{toxoMine: an integrated omics data warehouse for<i>Toxoplasma gondii</i>systems biology research}},
  booktitle = {Database},
  year = {2015},
  doi = {10.1093/database/bav066},
}
Figure