VisANT 4.0: Integrative network platform to connect genes, drugs, diseases and therapies

Zhenjun Hu, Yi-Chien Chang, Yan Wang, Chia-Ling Huang, Yang Liu, Feng Tian, Brian Granger, Charles DeLisiNucleic Acids Res 2013 · DOI

TL;DR

VisANT 4.0 is a powerful, updated web-based platform for visualizing and analyzing complex biological networks that integrate genes, drugs, diseases, and therapies. It addresses the challenge of understanding intricate disease mechanisms and drug actions by providing a unified framework for exploring these interconnected biological entities.

What Problem Does This Paper Address?

Modern biological research increasingly recognizes that diseases are not caused by single gene defects but by complex interactions within cellular networks. Similarly, drug actions and their efficacy are deeply intertwined with these biological networks. This complexity necessitates sophisticated tools that can integrate and visualize diverse biological data, including gene-gene interactions, drug-target relationships, and disease-therapy associations. Existing tools often focus on only one or two of these aspects, leaving a gap in providing a comprehensive, unified platform for exploring the multifaceted relationships between genes, drugs, diseases, and therapies. This paper aims to fill that gap by presenting VisANT 4.0, an enhanced platform designed for integrative network analysis.

Methods

VisANT 4.0 builds upon its previous versions by introducing several key advancements:

  • Metagraph-based Solution: The core of VisANT's approach is its metagraph model, which allows for the integration of hierarchical knowledge (like disease and therapy classifications) with adjacency relations (like protein-protein interactions). This enables the construction of 11 different types of networks.
  • Integrated Hierarchies: The platform now includes integrated search and navigation for disease (ICD-10) and drug (ATC) hierarchies. This allows users to explore these classifications in a structured manner.
  • Association Integration: VisANT 4.0 seamlessly integrates disease-gene, therapy-drug, and drug-target associations. This facilitates the construction and filtering of networks by linking these entities.
  • Annotation and Prediction: New functionalities allow for the annotation of genes and drugs with disease/therapy information. Furthermore, it enables the prediction of associated diseases/therapies for a given set of genes/drugs using enrichment analysis.
  • Network Transformation: The platform supports versatile network construction through various transformation functions, allowing users to create specific network views (e.g., disease-disease, therapy-drug, co-gene networks).
  • Enhanced User Interface: VisANT 4.0 features a revamped GUI with docking windows, providing a more flexible and customizable workspace. This includes easy customization of node and edge properties and a built-in legend node for clear visualization of different entity types.

Datasets

VisANT 4.0 integrates data from several established biological databases:

  • Disease Hierarchy: ICD-10 (International Classification of Disease) from the WHO.
  • Therapy Hierarchy: ATC (Anatomical Therapeutic Chemical) classification from the WHO.
  • Disease-Gene Associations: KEGG Disease database, OMIM, PharmGKB, and GAD (Genetic Association Database).
  • Therapy-Drug and Drug-Target Associations: KEGG Drug database and DrugBank.

The paper notes that data integration and synchronization procedures are employed to keep the information up-to-date.

Key Insights and Ideas

  • The Power of Integration: The most significant insight is the realization that understanding complex biological phenomena like disease mechanisms and drug actions requires an integrated approach that connects genes, drugs, diseases, and therapies within a unified network framework.
  • Hierarchical Knowledge is Crucial: Incorporating hierarchical classifications (like disease and drug categories) alongside interaction data is essential for building comprehensive and interpretable biological networks.
  • Network-Based Drug Discovery: The platform supports novel drug design strategies like drug repositioning, multi-target drugs, and combination therapy by enabling the visualization and analysis of drug-target and drug-disease relationships.
  • User-Centric Design: The emphasis on an enhanced GUI with docking windows and customizable properties highlights the importance of user experience in complex data visualization and analysis tools.

Key Results

The paper presents VisANT 4.0 as a functional and enhanced platform. While specific quantitative results are not detailed in terms of performance metrics (e.g., speedup), the paper demonstrates the platform's capability through:

  • Versatile Network Construction: The ability to construct 11 different types of networks, including meta-networks, disease-disease, therapy-therapy, disease-gene, therapy-drug, and complex multi-entity networks (e.g., disease-gene-drug-therapy).
  • Illustrative Examples: Figure 1 visually showcases the diverse network types that can be generated. Figure 2 demonstrates the hierarchy explorer and exploratory navigation. Figure 3 presents a drug-target network with a case study example.
  • Case Study: A case study on identifying repositioning drug candidates for breast cancer is presented, showcasing how VisANT can be used to filter networks based on disease associations and identify potential drug candidates. This study found 32 overlaps with a previous repositioning study.

Limitations and Future Work

  • Data Standardization: The paper acknowledges that while ICD-10 and ATC are standard classifications, not all disease databases fully support ICD-10. They mention a plan to merge disease hierarchies under ICD-10 as more databases adopt it.
  • Gene Name Search: The search box in the Hierarchy Explorer does not support gene name searches; this functionality is available in the VisANT toolbox.
  • Future Directions: The paper implies that the continuous development of VisANT will likely involve further integration of new data sources, refinement of analytical tools, and expansion of network types to address emerging challenges in systems biology and translational science. The platform's ability to accept various input formats (node/edge-lists, PSI-MI, GML) and save networks in multiple formats suggests a focus on interoperability and dissemination of research findings.

Summary updated

BibTeX
@article{hu2013visant,
  author = {Zhenjun Hu and Yi-Chien Chang and Yan Wang and Chia-Ling Huang and Yang Liu and Feng Tian and Brian Granger and Charles DeLisi},
  title = {{VisANT 4.0: Integrative network platform to connect genes, drugs, diseases and therapies}},
  journal = {Nucleic Acids Res.},
  year = {2013},
  doi = {10.1093/nar/gkt401},
}
Figure