High throughput ANI analysis of 90K prokaryotic genomes reveals clear species boundaries

Abstract

A fundamental question in microbiology is whether there is continuum of genetic diversity among genomes, or clear species boundaries prevail instead. Whole-genome similarity metrics such as Average Nucleotide Identity (ANI) help address this question by facilitating high resolution taxonomic analysis of thousands of genomes from diverse phylogenetic lineages. To scale to available genomes and beyond, we present FastANI, a new method to estimate ANI using alignment-free approximate sequence mapping. FastANI is accurate for both finished and draft genomes, and is up to three orders of magnitude faster compared to alignment-based approaches. We leverage FastANI to compute pairwise ANI values among all prokaryotic genomes available in the NCBI database. Our results reveal clear genetic discontinuity, with 99.8% of the total 8 billion genome pairs analyzed conforming to >95% intra-species and <83% inter-species ANI values. This discontinuity is manifested with or without the most frequently sequenced species, and is robust to historic additions in the genome databases.

Fig. 1

Correlation of FastANI and Mash-based ANI output with ANI_b values for datasets D1–D5. Because FastANI assumes a probabilistic identity cutoff that is set to 80% by default, it reports 76, 570, 4271, 464, and 130 genome matches for the individual queries in datasets D1–D5 respectively. To enable a direct quality comparison against FastANI, Mash is executed for only those pairs that are reported by FastANI. Notice that each dataset encompasses a different nucleotide identity range (x-axes). Gray line represents a straight line y = x plot for reference. Pearson correlation coefficients corresponding to these plots are listed separately in Table 2. Last plot shows error of these methods w.r.t. ANI_b using all five datasets

Fig. 2

Scaling up FastANI’s performance using multi-core parallel execution. We executed parallel FastANI processes on 40 physical cores, where each process was assigned an equally sized part of the reference D1–D5 databases. Left and right plots evaluate FastANI’s compute and indexing phase, respectively. FastANI achieves reasonable speedups on all datasets except the compute phase in D1 and D5, as their runtime on a single core is too small to begin with (Table 3)

Fig. 3

Genetic discontinuity observed using 90 K genomes. a Histogram plot showing the distribution of ANI values among the 90 K genomes. Only ANI values in the 76–100% range are shown. Out of total 8.01 billion pairwise genome comparisons, FastANI reported only 17M ANI values (0.21%) with ANI between 83 and 95% indicating a genetic discontinuum. Multiple colors are used to show how genomes from different genera are contributing to this distribution. Few peaks in the histogram arise from genera that have been extensively sequenced and dominate the database. b Density curves of ANI values in the ANI range 76–100%. Each curve shows the density curve corresponding to the database at a particular time period. Discontinuity in all four curves is observed consistently. c. Distribution of ANI values with each comparison labeled by the nomenclature of genomes being compared. All the comparisons between Escherichia coli and Shigella spp. have been labeled separately. The 95% ANI threshold on x-axis serves as a valid classifier for comparisons belonging to same and different species

Fig. 4

FastANI algorithm explained using synthetic and real examples. a Illustration of FastANI’s work-flow for computing ANI between a query genome and a reference genome. Five mappings are obtained from three query fragments using Mashmap. M_forward saves the maximum identity mapping for each query fragment. In this example, M_forward = {m₂, m₄, m₅}. From this set, M_reciprocal picks m₄ and m₅ as the maximum identity mapping for each reference bin. Mapping identities of orthologous mappings, thus found in M_reciprocal, are finally averaged to compute ANI. b FastANI supports visualization of the orthologous mappings M_reciprocal that are used to estimate the ANI value using genoPlotR. In this figure, ANI is computed between Bartonella quintana strain (NC_018533.1) as query and Bartonella henselae strain (NC_005956.1) as reference. Red line segments denote the orthologous mappings computed by FastANI for ANI estimation