StrainSelect

StrainSelect provides a strain-resolved microbiome reference database that disambiguates genome assemblies, links DNA biomarkers to procurable strains, and unifies taxonomic references across shotgun metagenomics and 16S rRNA gene data.

Key Features:

• Knowledge graph disambiguation: A knowledge graph represents contiguous DNA sequences, genome assemblies, strain monikers, and bio-resource center (BRC) catalog numbers with edges added only for synonyms or direct derivatives to distinguish strains from assemblies.
• Extensive strain collection: The database catalogs 681,087 strains and addresses replicate RefSeq genome assemblies while organizing each strain into eight taxonomic ranks with bootstrap confidence inversely correlated to genome assembly contamination.
• Metagenomic integration: Contains 636,568 strains with at least one 16S rRNA gene, 245,005 strains with annotated genome assemblies, and 36,671 strains procurable from at least one BRC for combined shotgun and amplicon analyses.
• Unified taxonomic reference: Provides a single taxonomic reference usable for taxonomic and functional assignments in both 16S rRNA amplicon and shotgun metagenomics workflows.
• Exported computational formats: Outputs include igraph and tabular vertex-edge formats compatible with Neo4J and dereplicated MinHash and fasta databases for use with sourmash and usearch pipelines.

Scientific Applications:

• Disease biomarker discovery: Enables strain-resolved identification of DNA biomarkers from metagenomic and 16S rRNA data.
• Therapeutic and experimental validation: Facilitates mapping of DNA biomarkers to laboratory-procurable strains via BRC catalog links for experimental follow-up.
• Taxonomic and functional profiling: Serves as a reference for taxonomic and functional assignment across shotgun metagenomics and amplicon sequencing studies.
• Strain procurement mapping: Supports procurement-focused analyses by linking strain identifiers to BRC catalog numbers.

Methodology:

Constructed a knowledge graph with vertices for contiguous DNA sequences, genome assemblies, strain monikers, and BRC catalog numbers and edges only for synonyms or direct derivatives; assigned strains to eight taxonomic ranks with bootstrap confidence metrics; produced igraph and tabular vertex-edge outputs and dereplicated MinHash and fasta databases for sourmash and usearch.