GdbMTB: a manually curated genomic database of magnetotactic bacteria

Runjia Ji, Yongxin Pan, Wei LinDatabase J. Biol. Databases Curation 2026 · DOI

TL;DR

Magnetotactic bacteria (MTB) are fascinating microorganisms that navigate using Earth's magnetic field, but their genomic data is scattered and inconsistent. This paper introduces GdbMTB, the first comprehensive, manually curated genomic database for MTB, integrating 365 genomes and rich metadata to facilitate research into their unique biology and ecological roles.

What Problem Does This Paper Address?

Magnetotactic bacteria (MTB) are a unique group of microorganisms that possess intracellular magnetic crystals called magnetosomes, allowing them to orient themselves along geomagnetic field lines. This ability is crucial for their survival and navigation in diverse aquatic environments. While genomic sequencing has revolutionized our understanding of these microbes, the publicly available MTB genomic data is fragmented across multiple databases, often with varying quality and incomplete contextual information (like sampling location and environmental conditions). This makes it challenging for researchers to conduct comprehensive comparative analyses, explore their evolutionary history, or understand their ecological functions. GdbMTB aims to solve this by consolidating and standardizing this dispersed data into a single, user-friendly resource.

Methods

The creation of GdbMTB involved several key methodological steps:

  • Data Collection: MTB genomes were gathered from major public databases including NCBI, eLMSG, IMG, and RAST. A standardized naming convention was applied using accession numbers.
  • Metadata Curation: Crucial metadata was collected from both public databases and original scientific publications. This included:
  • Sample Metadata: Environmental context such as geographic coordinates, sampling depth, salinity, temperature, and oxygen concentration. Habitats were classified into 11 categories.
  • Genome Metadata: Information related to genome submission, including accession identifiers, sequencing platform, submitter institution, and submission date.
  • Reference Metadata: Details of associated publications, such as authors, titles, journals, PMIDs, and DOIs.
  • Genome Quality Assessment: A suite of established bioinformatics tools was employed to assess genome quality. This included:
  • CheckM and CheckM2 for completeness and contamination estimation.
  • BUSCO for assessing gene content completeness.
  • GUNC for detecting chimerism and contamination.
  • QUAST for evaluating general genomic features like contig count, N50, and genome size.
  • Prodigal for coding sequence (CDS) prediction.
  • rRNAFinder and Aragorn for identifying ribosomal and transfer RNA genes.
  • Taxonomic Classification and Phylogeny:
    • GTDB-tk was used for taxonomic assignment.
  • dRep was employed for dereplication at the species level (95% average nucleotide identity threshold).
  • A phylogenomic tree of MTB was constructed using 120 single-copy marker proteins identified by GTDB-tk, aligned with MAFFT and trimmed with trimAL, and then phylogenetically analyzed using IQ-TREE.
  • Magnetosome Gene Cluster (MGC) Annotation: MagCluster was used to annotate MGCs within the genomes. Representative MGCs were manually curated and visualized.
  • Website Development: The GdbMTB website was built using Python 3 and the Flask framework, with a front-end developed in HTML, CSS, and JavaScript. It is deployed on PythonAnywhere and designed for compatibility across major browsers.

Datasets

The primary dataset used in this study is a compilation of 365 publicly available MTB genomes and their associated metadata, collected as of June 2024. These genomes were sourced from various public repositories. The paper also references the use of established databases for taxonomic classification (Genome Taxonomy database release 220) and for identifying conserved genes (e.g., 120 single-copy marker proteins). No new benchmark datasets were introduced; instead, the contribution is the curated integration of existing genomic data.

Key Insights and Ideas

  • Consolidation is Key: The most significant contribution is the creation of a single, unified resource for MTB genomic data. This addresses the critical bottleneck of scattered and inconsistent information, making MTB research more accessible and efficient.
  • Beyond Just Genomes: The emphasis on rich, manually curated metadata (environmental context, publication details) is crucial. This allows researchers to connect genomic information with ecological and evolutionary insights, moving beyond simple sequence analysis.
  • Quality Matters: The rigorous quality assessment of each genome is vital. By classifying genomes into "high," "medium," and "low" quality categories, GdbMTB provides users with transparency and helps prevent downstream analyses from being skewed by poor-quality data.
  • Diversity Revealed: The database highlights the broad phylogenetic distribution of MTB across 16 bacterial phyla, underscoring the widespread nature of magnetotaxis. The classification into "confirmed MTB" (cMTB) and "potential MTB" (pMTB) based on evidence of magnetotactic behavior or MGC presence is a useful distinction for future research.
  • A Foundation for Future Discovery: GdbMTB is not just a static repository; it's presented as a platform that will be continuously updated and enhanced with new data and analytical capabilities, supporting ongoing and future MTB research.

Key Results

  • 365 MTB Genomes Integrated: GdbMTB currently hosts 365 publicly available MTB genomes.
  • Diverse Taxonomic Representation: These genomes span 16 bacterial phyla, 35 classes, and 49 orders. Pseudomonadota is the most represented phylum, followed by Desulfobacterota and Nitrospirota.
  • Quality Distribution: Of the 365 genomes, 176 are classified as high-quality (≥90% completeness, ≤5% contamination), 179 as medium-quality (≥50% completeness, ≤10% contamination), 4 as low-quality, and 6 as high-contamination.
  • Phylogenetic Insights: A species-level phylogenomic tree was constructed, encompassing 252 unique species-level representative MTB genomes and 278 non-MTB relatives, delineating genetic relationships.
  • Global and Environmental Coverage: The database includes MTB genomes from six continents and various habitat types, with a notable representation from freshwater environments.
  • Historical Data: The compilation covers genomes published from 2005 to 2024, illustrating the growth of MTB genomic research over time.

Limitations and Future Work

  • Limitations:
    • Culturing Bias: The paper acknowledges that the majority of MTB remain uncultured, which inherently limits the depth of functional and physiological characterization that can be directly inferred from genomic data alone.
  • Data Quality Variability: While GdbMTB provides quality assessments, the original data quality from public repositories can still vary, and some genomes are flagged with high contamination.
  • Experimental Validation: For "potential MTB" (pMTB) genomes identified solely by MGCs, experimental validation of magnetotactic behavior is often lacking.
  • Future Work:
    • Continuous Updates: The database will be continuously updated with new MTB genome data as it becomes available.
  • Enhanced Online Capabilities: Future work will focus on improving online analysis capabilities, making data exploration even more intuitive and user-friendly.
  • Integration with Other Databases: The paper mentions providing links to external databases, suggesting a continued effort to facilitate cross-database exploration.
  • Further Functional Analysis: While MGCs are annotated, future work could expand on metabolic function profiling and comparative genomics to better understand ecological roles.

Summary updated

BibTeX
@article{ji2026gdbmtb,
  author = {Runjia Ji and Yongxin Pan and Wei Lin},
  title = {{GdbMTB: a manually curated genomic database of magnetotactic bacteria}},
  journal = {Database J. Biol. Databases Curation},
  year = {2026},
  doi = {10.1093/database/baaf090},
}
Figure