What this entity is
Canonical name: DBSCAN. 1
Known aliases: None reported in these documents. 1,2
Stable external identifier: None reported in these documents. 1,2
DBSCAN is a density-based clustering method used after umap (uniform manifold approximation and projection) to identify distinct antimicrobial-resistance (AMR) ecotypes. 1 The AlphaEarth Embeddings, Geography & Environment Explorer shows that the same method was also applied to UMAP projections of 64-dimensional AlphaEarth embeddings, where eps=0.5 produced 320 embedding clusters. 2
Key facts across projects
In the AMR ecotype analysis, DBSCAN included 974 species with at least 15 genomes. 1 Of these, 190 (19.5%) formed at least 2 distinct AMR ecotypes. 1 The median silhouette score for the ecotype clusters was 0.620. 1 The analysis generated 176,177 ecotype assignments and 974 ecotype summaries. 1
Environment–ecotype association testing was limited because 52.7% of genomes had no classifiable isolation source, and only 2 species had sufficient within-species environmental diversity for chi-squared testing after strict expected-frequency criteria. 1 Case-study UMAP plots for klebsiella-pneumoniae, staphylococcus-aureus, and salmonella-enterica showed visible environmental structuring, but the underpowered statistical tests did not establish a general environment–ecotype association. 1
The AlphaEarth application supports DBSCAN as a reusable clustering step for embedding analysis: many of its 320 clusters were dominated by a single harmonized environment category, while Human gut and Human clinical genomes were distributed across many clusters. 2 This refines the AMR use case by showing that apparent clusters can reflect environmental and geographic structure, not only AMR ecotypes. 2 The report cautions that DBSCAN cluster structure depends on UMAP parameters and that eps=0.5 may produce overly fine-grained clusters; coarser clustering may better match environment categories. 2
DBSCAN therefore contributes to the environmental-resistome analysis by providing the clustering step for testing whether within-species AMR repertoires form ecologically structured groups. 1
Related pages
- umap — dimensionality-reduction method used before DBSCAN in both the AMR ecotype and AlphaEarth embedding analyses. 1,2
- Within-Species AMR Strain Variation — source report describing the AMR ecotype analysis. 1
- AlphaEarth Embeddings, Geography & Environment Explorer — source report describing DBSCAN clustering of AlphaEarth embeddings. 2
- environmental-resistome — cross-project concept covering environmental structure in AMR repertoires. 1
- environment-embedding-geography — concept covering geographic and environmental structure in AlphaEarth embeddings. 2