ViMIC: a database of human disease-related virus mutations, integration sites and cis-effects

Ying Wang, Yuantao Tong, Zeyu Zhang, Rongbin Zheng, Danqi Huang, Jinxuan Yang, Hui Zong, Fanglin Tan, Yujia Xie, Honglian Huang, Xiaoyan ZhangNucleic Acids Res 2022 · DOI

TL;DR

ViMIC is a novel, comprehensive database that integrates human disease-related virus mutations (VMs), viral integration sites (VISs), and their associated cis-regulatory effects on host genes. This resource aims to bridge the gap in scattered and unstandardized data, providing a valuable tool for researchers to explore the complex mechanisms of virus-induced diseases.

What Problem Does This Paper Address?

Viruses can cause serious diseases through various mechanisms, including mutations within their own genomes and integration of their genetic material into the host DNA. While a wealth of research exists on these topics, the data is often fragmented across numerous publications, lacking standardization and easy accessibility. This makes it challenging for researchers to gain a holistic understanding of virus-host interactions, identify patterns, and develop new therapeutic strategies. Existing databases tend to focus on either viral mutations or integration sites, but rarely integrate both with functional cis-regulatory information. ViMIC aims to address this by creating a unified platform that consolidates and curates information on virus mutations, integration sites, and their impact on host gene regulation.

Methods

ViMIC employs a multi-pronged approach to data acquisition and processing:

  • Text Mining and Manual Curation: A sophisticated text-mining system, utilizing Entrez Programming Utilities and Python scripts, was developed to extract relevant information (mutations, integration sites, target genes, diseases) from millions of PubMed and PMC articles. This automated extraction is followed by rigorous manual curation to ensure data accuracy and standardization.
  • Data Integration: Data from various public repositories, including NCBI Nucleotide, GEO, Cistrome Data Browser, and VISDB, are integrated into ViMIC.
  • VIS-Cistrome Overlap Analysis: A key feature is the analysis of overlaps between viral integration sites (VISs) and host cistrome data (histone modifications, transcription factor binding sites, and chromatin accessibility). This is achieved by building GIGGLE indexes for peak files and using them to identify significant genomic loci overlaps.
  • Correlation Analysis: For target genes, ViMIC performs correlation analysis between gene expression levels and the fraction of infiltrating immune cells, using CIBERSORT for estimation.
  • GRCh38 Conversion: All genomic coordinates are converted to the GRCh38 human reference genome for standardization.

Datasets

ViMIC integrates data from several public sources:

  • PubMed and PMC: For text mining of virus-related articles.
  • NCBI Nucleotide Database: For viral sequence information.
  • Cistrome Data Browser: For human and mouse ChIP-seq, DNase-seq, and ATAC-seq data.
  • VISDB: For curated viral integration sites.
  • GEO (Gene Expression Omnibus) and TCGA (The Cancer Genome Atlas): For gene expression datasets of virus-infected diseases.

The database itself contains:

  • 31,712 Virus Mutation (VM) entries.
  • 1,110,015 Virus Sequence entries.
  • 78 Histone Mark types overlapped with VISs.
  • 1,358 Transcription Regulator types overlapped with VISs.
  • 95,280 Viral Integration Sites (VISs) with overlaps.
  • 16,310 Target Genes.
  • 77 Virus-Related Diseases.
  • 127 Clinical Annotation and Data entries.

Key Insights and Ideas

  • The Power of Integration: The core idea is that understanding virus-host interactions requires integrating diverse data types – mutations, integration events, and their impact on host gene regulation. ViMIC demonstrates the feasibility and value of such integration.
  • Functional Annotation of VISs: By overlapping VISs with cistrome data, ViMIC provides crucial insights into whether viral integration occurs in functional regions of the host genome, potentially influencing gene expression.
  • Bridging the Gap: ViMIC directly addresses the problem of scattered and unstandardized data, making it easier for researchers to access and utilize information that was previously difficult to find and compare.
  • Cis-Effects are Crucial: The database highlights the importance of "cis-effects" – how viral integration can directly influence the regulation of nearby host genes, impacting disease progression and immune response.

Key Results

  • ViMIC provides a comprehensive resource containing information on 8 viruses, 77 human diseases, 31,712 virus mutations, and 105,624 viral integration sites.
  • The database allows users to explore the cis-regulatory landscape around viral integration sites, including overlaps with 78 histone modifications and binding of 1,358 transcription regulators.
  • For target genes, ViMIC offers correlation analysis between gene expression and immune cell infiltration, providing a deeper understanding of disease mechanisms.
  • The web interface is designed for user-friendly browsing, searching, and visualization, with examples demonstrating how to explore virus mutation sites, viral integration sites, and target gene information.

Limitations and Future Work

  • Limited Virus Coverage (Initial): The current version focuses on 8 specific viruses. Future work will involve adding more viruses to expand the database's utility.
  • Automated Mutation Recognition: The authors plan to further optimize the technology for automatically recognizing mutations from literature.
  • In-depth Sequence Analysis: The integration of sequences around VISs will enable more in-depth analysis of functional roles in virus-host interactions.
  • Batch Upload and Analysis: The upload module will be enhanced for batch processing of potential VISs and performing online analysis.
  • Additional Immune Deconvolution Tools: Integration of tools like TIMER and xCell is planned to improve the analysis of gene expression and immune cell infiltration correlations.
  • Continuous Maintenance and Updates: ViMIC will be continuously maintained and updated as new data on VMs, VISs, and cistrome data become available.

Summary updated

BibTeX
@article{wang2022vimic,
  author = {Ying Wang and Yuantao Tong and Zeyu Zhang and Rongbin Zheng and Danqi Huang and Jinxuan Yang and Hui Zong and Fanglin Tan and Yujia Xie and Honglian Huang and Xiaoyan Zhang},
  title = {{ViMIC: a database of human disease-related virus mutations, integration sites and cis-effects}},
  journal = {Nucleic Acids Res.},
  year = {2022},
  doi = {10.1093/nar/gkab779},
}
Figure