Fig. 1

Reconstructed sampling curves of 13 Accumulibacter genomes by the exhausted subsampling method. The number of genes was calculated as a function of adding an n^th genome into the (n-1) genomes. The total number (≤ TN) of effective permutation for each n was represented by the number of circles, which were obtained by different genome combinations. a Core-genome sampling curve. The average number of core genes at each n number was plotted as squares, and the continuous curve represented the least-squares fit of the function \( \mathrm{Fc}=\mathrm{Kc}\ \exp \left[-\frac{n}{\uptau \mathrm{c}}\right]+\Omega \). The best hit vector for Kc, τc, and Ω was 3000, 3.04, and 1761 with correlation r² 0.98. b Strain-specific genome sampling curve. The average number of strain-specific genes at each n number was plotted as squares, and the continuous curve represented the least-squares fit of the function \( \mathrm{Fs}=\mathrm{Ks}\ \exp \left[-\frac{n}{\uptau \mathrm{s}}\right]+\mathrm{tg}\left(\uptheta \right) \). The best hit vector for Ks, τs, and tg(θ) was 1234, 4.05, and 258 with correlation r² 1.00. c Pan-genome sampling curve. The average number of all genes (size of pan-genome) at each n number was plotted as squares, and the continuous curve represented the least-squares fit of the function \( P(n)=D+\mathrm{tg}\left(\uptheta \right)\left[n-1\right]+\mathrm{Ksexp}\left[-\frac{2}{\uptau \mathrm{s}}\right]\frac{1-\exp \left[-\frac{n-1}{\uptau \mathrm{s}}\right]}{1-\exp \left[-\frac{1}{\uptau \mathrm{s}}\right]} \). With the best hit vector 1234, 4.05, and 258 for Ks, τs, and tg(θ) of strain-specific genome fitting and D as 3634; the correlation r² of pan-genome fitting is 1.00