SilkDB 3.0: visualizing and exploring multiple levels of data for silkworm

Lu Fang, Zhaoyuan Wei, Yongjiang Luo, Hailong Guo, Guoqing Zhang, Qingyou Xia, Yi WangNucleic Acids Res 2020 · OpenReview

TL;DR

SilkDB 3.0 is a significantly upgraded, open-access database and visualization platform for the silkworm (Bombyx mori). It integrates a vast array of genomic, transcriptomic, and epigenomic data, offering an intuitive interface for researchers to explore complex biological relationships and accelerate discoveries in silkworm biology and related insect research.

What Problem Does This Paper Address?

The silkworm (Bombyx mori) is a crucial model organism for insect research and a valuable bioreactor for producing biomaterials. While previous versions of SilkDB and other databases have provided valuable silkworm data, the rapid generation of diverse "omics" data (genome, transcriptome, epigenome, etc.) over the past decade has outpaced the ability of existing resources to integrate, visualize, and analyze this information comprehensively. Researchers needed a unified platform that could handle multiple data levels, offer advanced visualization tools, and provide an improved user experience to facilitate deeper biological insights and hypothesis generation.

Methods

SilkDB 3.0 was built upon a robust technological stack, including a MySQL database management system, and a dynamic web interface developed with Python, HTML, CSS, JavaScript, and jQuery. The platform is hosted on an Ubuntu Linux machine with an Apache2 server.

Key methodological advancements include:

  • Genome Assembly and Annotation: The silkworm genome assembly was significantly improved using PacBio long reads and Hi-C data for chromosome-level scaffolding. This resulted in a more complete and accurate assembly compared to previous versions. Gene annotation was re-performed using MAKER, leading to a higher number of identified protein-coding genes.
  • Transcriptome Analysis: Extensive RNA-seq datasets from various silkworm tissues and developmental stages were processed using standard bioinformatics pipelines (fastp, HISAT2, SAMtools, ballgown) to generate gene expression values and coexpression networks.
  • Pangenome Analysis: A large collection of silkworm strains (163 in total) was analyzed to characterize genomic variation, including SNPs and indels, using BWA-MEM and BCFtools.
  • Hi-C Data Integration: Hi-C datasets were processed and visualized using HiGlass to enable the exploration of 3D chromatin interactions across multiple resolutions.
  • Visualization Tools: The platform incorporates a suite of visualization tools, including eFP (electronic fluorescent pictograph) for expression patterns, Cell eFP for subcellular localization, Coexpression network visualization, Gene Family trees, Pangenome viewers (MSA, JBrowse), and 3D structure viewers.

Datasets

SilkDB 3.0 integrates a wide range of datasets, including:

  • Genome Data: A high-quality, chromosome-level assembly of the silkworm genome.
  • Transcriptome Data: 253 RNA-seq datasets from various silkworm tissues and developmental stages.
  • Pangenome Data: Data from 163 silkworm strains, including both domestic and wild types.
  • Hi-C Data: Datasets from six silkworm tissue types to map chromatin interactions.
  • Comparative Genomics Data: Annotated genes from Aedes aegypti, Drosophila melanogaster, Spodoptera litura, Tribolium castaneum, and Trichoplusia ni for ortholog analysis.

No new benchmarks were introduced; instead, the paper focuses on integrating and presenting existing data in a novel, user-friendly manner.

Key Insights and Ideas

  • Unified Multi-Level Data Exploration: The core innovation is the seamless integration of diverse biological data types (genomic, transcriptomic, epigenomic) into a single, interactive platform. This allows researchers to move fluidly between different levels of biological organization, fostering a more holistic understanding of gene function and regulation.
  • Enhanced Genome Assembly and Annotation: The improved genome assembly and annotation provide a more accurate foundation for all subsequent analyses, increasing the reliability of the data presented.
  • Interactive Visualization for Complex Relationships: The suite of visualization tools (eFP, coexpression networks, Hi-C maps) transforms complex datasets into easily interpretable graphical representations, enabling the discovery of non-obvious patterns and relationships.
  • Facilitating Comparative Genomics: The inclusion of ortholog data from related insect species allows for comparative studies, aiding in the understanding of evolutionary processes and the identification of conserved or divergent gene functions.

Key Results

  • Improved Genome Assembly: SilkDB 3.0 features a high-quality, chromosome-level assembly of the silkworm genome (~468.3 Mb) with 16,069 protein-coding genes, a significant improvement over previous versions.
  • Comprehensive Data Integration: The database integrates a vast amount of data, including transcriptome data for gene expression analysis, pangenome data for genetic variation, and Hi-C data for chromatin interaction mapping.
  • User-Friendly Interface: The updated interface is designed to be simple and intuitive, allowing users to easily search for genes, explore functional modules, and visualize data at multiple biological levels.
  • Valuable Resource for Research: The database has already been cited over 400 times (Google Scholar) for its previous versions, indicating its significant utility. SilkDB 3.0 is expected to further enhance its value to the silkworm and broader insect research community.

Limitations and Future Work

  • Continuous Improvement of Assembly and Annotation: The authors acknowledge that they will continue to refine the quality of the silkworm genome assembly and annotations.
  • Augmenting Data Types: Future plans include expanding the database with new types of biological data, such as phenotype and metabonomics data.
  • Expanding Comparative Genomics: The team aims to incorporate other lepidopteran genomes into SilkDB and develop new comparative genomics visualization tools for these additional genomes.
  • Data Accessibility: While the platform is open-access, the sheer volume of data and the complexity of some visualizations might still present a learning curve for novice users.

Summary updated

BibTeX
@inproceedings{fang2020silkdb,
  author = {Lu Fang and Zhaoyuan Wei and Yongjiang Luo and Hailong Guo and Guoqing Zhang and Qingyou Xia and Yi Wang},
  title = {{SilkDB 3.0: visualizing and exploring multiple levels of data for silkworm}},
  booktitle = {Nucleic Acids Res},
  year = {2020},
  url = {https://openreview.net/forum?id=EKxeWsdnC7},
}
Figure