What this entity is
Canonical name: Average nucleotide identity. Known alias: ANI. ANI is a genome-comparison method represented in these reports as a distance matrix for comparing phylogenetic relatedness among strains and species. 1
Stable external identifier: No stable external identifier is reported in the source documents. 1
Key facts
The study compared ANI distance matrices with antimicrobial-resistance (AMR) gene Jaccard-distance matrices across 1,261 species using a Mantel test. 1
Of the 1,261 species tested, 701 (55.6%) showed significant phylogenetic signal at FDR < 0.05, where FDR is the false-discovery rate. 1
The median Mantel r for all AMR genes was 0.247, and 87.8% of species showed a positive correlation between ANI distance and AMR Jaccard distance. 1
Non-core AMR genes had a median Mantel r of 0.222, compared with 0.117 for core AMR genes; the paired t-test result was t = -8.35, p = 7.0e-16, n = 489. 1
The report interprets the stronger signal for non-core AMR genes as supporting the hypothesis that acquired resistance elements can become stably maintained and vertically transmitted within lineages, while noting that near-universal core genes have little Jaccard-distance variance and can therefore produce lower distance-based Mantel correlations independently of biology. 1
ANI extraction for the Mantel analyses was limited to species with <=500 genomes. 1 The study generated 1,259 ANI matrices. 1 The <=500-genome limit excluded mega-species such as Escherichia coli, which had 15,388 genomes, and Klebsiella pneumoniae, which had 14,240 genomes, from the phylogenetic-signal analysis. 1
The co-fitness/pangenome study refines ANI’s use as a relatedness measure by reporting near-clonal organisms at ANI 99.47% for Ddia6719 and ANI 99.66% for pseudo3_N2E3; both retained enough accessory-gene variation to detect co-inheritance. 2 It also found higher cofit-pair phi among near genomes (mean = 0.102) than medium-distance genomes (mean = 0.067), supporting a shared-ancestry contribution while noting that limited far-distance representation prevented full separation of phylogenetic and functional effects. 2
The ecotype analysis supports ANI-based phylogenetic distance as a genome-wide relatedness measure: among 172 species, the median partial correlation—association between two variables while accounting for another variable—was 0.0143 for phylogeny versus 0.0025 for environment, and phylogeny dominated in 60.5% of species. 3 This result is reported in the context of 13,381 genomes across 224 species, with correlation results available for 172 species. 3
The environmental reanalysis supports the original ecotype null conclusion within a consistent method: ANI distances were combined with AlphaEarth embeddings and gene-cluster memberships, and environmental species did not have stronger environment–gene-content partial correlations than human-associated species. 4 Environmental species had median partial correlation 0.051 versus 0.084 for human-associated species; the one-sided Mann–Whitney U test gave U=1536 and p=0.83. 4 The reanalysis refines the original result by showing that the clinical sampling bias was real but did not explain the weak signal, while its full-genome, non-downsampled extraction produced an all-species median of 0.081 versus 0.003 originally; these absolute values are not comparable across methodologies. 4
Relationships
ANI supplied the phylogenetic-distance component for the study’s analysis of environmentally and lineage-structured AMR variation. 1
It also connects the AMR analysis to pangenome integration, because ANI-based relatedness was compared with within-species AMR gene-repertoire differences. 1 The method is further used in the co-fitness analysis linking pangenome co-occurrence with phylogenetic distance; details are in Co-fitness and Co-inheritance in Bacterial Pangenomes. 2
The ecotype analysis further links ANI-derived phylogenetic distances to environmental-versus-ancestry tests of pangenome gene-content similarity; details are in Ecotype Correlation Analysis. 3 The reanalysis extends that link by integrating ANI distances with genome-level environment classifications and AlphaEarth embeddings; details are in Ecotype Reanalysis — Environmental vs Human-Associated Species. 4
Limitations
The report cautions that the stronger phylogenetic signal for non-core AMR genes may be partly statistical because core genes are nearly universal and consequently provide little Jaccard-distance variance. 1
The report proposes subsampling strategies for species with more than 500 genomes to extend ANI-based Mantel analyses to larger species collections. 1 The co-fitness study likewise had limited phylogenetic control: stratification was available for 7 of 9 organisms, and most lacked genomes in the far stratum (>0.05 branch distance). 2
The ecotype analysis refines this interpretation by finding no significant environmental effect in 156 of 172 species (90.7%), but it also shows that the result is constrained by AlphaEarth embeddings covering only 28.4% of genomes and by partial correlations that assume linear relationships between distance matrices. 3 Thus, its weak environmental signal does not exclude effects on specific gene subsets or ecological effects not captured by the environmental representation. 3 The reanalysis adds that environmental species had a 21% NaN-correlation rate versus 7% for human-associated species and that a direct downsampled-versus-full-genome comparison is needed to explain the methodological discrepancy; its within-method group comparison remains valid despite the non-comparable absolute correlations. 4