NDEx IQuery: a multi-method network gene set analysis leveraging the Network Data Exchange

Rudolf Pillich, Jake Y. Chen, Christopher Churas, Daniel Tik-Pui Fong, Trey Ideker, Sn Liu, Benjamin M. Gyori, Klas Karis, Keiichiro Ono, Alexander R. Pico, Dexter PrattbioRxiv 2022 · DOI

Abstract

Motivation: The investigation of sets of genes using biological pathways is a common task for researchers and is supported by a wide variety of software tools. This type of analysis generates hypotheses about the biological processes active or modulated in a specific experimental context. Results: The NDEx Integrated Query (IQuery) is a new tool for network and pathway-based gene set interpretation that complements or extends existing resources. It combines novel sources of pathways, integration with Cytoscape, and the ability to store and share analysis results. The IQuery web application performs multiple gene set analyses based on diverse pathways and networks stored in NDEx. These include curated pathways from WikiPathways and SIGNOR, published pathway figures from the last 27 years, machine-assembled networks using the INDRA system, and the new NCI-PID v2.0, an updated version of the popular NCI Pathway Interaction Database. IQuery’s integration with MSigDB and cBioPortal now provides pathway analysis in the context of these two resources. Availability and Implementation: IQuery is available at https://www.ndexbio.org/iquery and is implemented in Javascript and Java. Contact: Dexter Pratt ([email protected]).

Introduction

In modern genomics, researchers frequently generate lists of genes—"gene sets"—and must determine the biological processes they represent. While numerous tools exist for this purpose, the current landscape is hindered by several challenges. Data silos prevent the aggregation of molecular interaction sources, forcing researchers to navigate multiple platforms. Furthermore, when databases do aggregate data, they often employ rigid, "normalized" schemas that strip away the original structure and nuances of the source data. Analysis results are typically published as static images, rendering them non-computable and difficult to re-analyze. Finally, moving data between analysis platforms (e.g., MSigDB) and visualization tools (e.g., Cytoscape) remains a cumbersome process.

Methods and Implementation

IQuery is designed as a modular, service-oriented application utilizing a Javascript/React frontend and a Java backend. Central to its functionality is the Cytoscape Exchange (CX) format, a flexible JSON-based structure that enables the transmission of complex networks while preserving biological meaning and visual styling.

To facilitate the identification of relevant pathways, IQuery implements three ranking metrics:

  1. Similarity: Employs a Cosine similarity score based on TF-IDF (Term Frequency-Inverse Document Frequency). This ensures that rare, specific genes shared between a query and a network carry more weight than common genes.
  2. P-Value: A standard hypergeometric test adjusted for false discovery rates (Benjamini-Hochberg), measuring the statistical probability of overlap occurring by chance.
  3. Overlap: A simple count of shared genes, useful for stable sorting when combined with p-values.

Data Resources

IQuery leverages the Network Data Exchange (NDEx) to provide access to four primary categories of network data:

  • Curated Pathways: Includes 675 human pathways from WikiPathways, 90 models from SIGNOR, and the updated NCI-PID v2.0.
  • Pathway Figures: A unique collection of over 30,000 gene sets extracted from 25 years of published literature using Optical Character Recognition (OCR) and Named Entity Recognition (NER).
  • INDRA GO: Over 6,000 networks generated by the INDRA system, which uses automated literature mining to extract high-confidence interactions related to specific Gene Ontology (GO) terms.
  • Interactomes: Access to 20 large-scale interaction networks, such as BioPlex 3 and BioGRID, for on-demand queries against protein-protein interaction sets.

Results

The IQuery interface provides a unified environment for multi-method analysis. As illustrated in Figure 1, the tool offers immediate visual feedback upon gene set entry.

The IQuery interface showing the Curated Pathways tab. The UI includes an input box for gene sets (red), analysis tabs (green), a results panel with ranking metrics (blue), and a graphical rendering module (orange).

The interface includes an input box for gene sets, multiple analysis tabs, a results panel with ranking metrics, and a graphical rendering module. IQuery automatically handles gene name normalization, highlighting valid genes and striking through invalid ones. Users can toggle between ranking methods to observe how their gene set aligns with known biology. A key advantage is the "computable sharing" capability; because IQuery is integrated with NDEx, users can save analysis results directly to a cloud account as live networks, which can then be opened in Cytoscape for further editing or shared via URL.

Discussion and Future Work

IQuery represents a significant advancement by transforming static pathway data into interactive, computable networks. It effectively "resurrects" data from the literature graveyard through OCR and utilizes machine-augmented curation to keep resources like NCI-PID up-to-date.

While powerful, the authors note that ranking results should be interpreted as general guidance due to variations in source granularity. Future developments will focus on expanding community contributions, enabling private instances for proprietary data, and deepening integration with other bioinformatics applications to further streamline research workflows.

Summary updated

Abstract

Abstract Motivation The investigation of sets of genes using biological pathways is a common task for researchers and is supported by a wide variety of software tools. This type of analysis generates hypotheses about the biological processes active or modulated in a specific experimental context. Results The NDEx Integrated Query (IQuery) is a new tool for network and pathway-based gene set interpretation that complements or extends existing resources. It combines novel sources of pathways, integration with Cytoscape, and the ability to store and share analysis results. The IQuery web application performs multiple gene set analyses based on diverse pathways and networks stored in NDEx. These include curated pathways from WikiPathways and SIGNOR, published pathway figures from the last 27 years, machine-assembled networks using the INDRA system, and the new NCI-PID v2.0, an updated version of the popular NCI Pathway Interaction Database. IQuery’s integration with MSigDB and cBioPortal now provides pathway analysis in the context of these two resources. Availability and Implementation IQuery is available at https://www.ndexbio.org/iquery and is implemented in Javascript and Java. Contact Dexter Pratt ( [email protected] )
BibTeX
@misc{pillich2022ndex,
  author = {Rudolf Pillich and Jake Y. Chen and Christopher Churas and Daniel Tik-Pui Fong and Trey Ideker and SN Liu and Benjamin M. Gyori and Klas Karis and Keiichiro Ono and Alexander R. Pico and Dexter Pratt},
  title = {{NDEx IQuery: a multi-method network gene set analysis leveraging the Network Data Exchange}},
  year = {2022},
  doi = {10.1101/2022.10.24.513552},
  howpublished = {bioRxiv},
}
Figure