Fig. 4: Potato pan-genome analysis.
a, Estimating the total amount of sequence variation in the European potato. Pan-genome graphs with different amounts of haplotype genomes (that is, x from 1 to 40) were used to fit a curve (that is, f(x)) that revealed the estimated size of the complete potato pan-genome. The unit of f(x) in the given model is Mb; for example, f(1) = 729.3 Mb. The sequenced genomes cover 85% of this estimated variation. b, Pan- and core-genomes of the European potatoes (labelled as Tetraploid) compared with the pan- and core-genomes of diploid potato samples from recent studies22 (labelled as Diploid). The number of genes shared by all haplotype genomes (core-genome: 5,602; 11.63%) in tetraploid potatoes was less than half of what was found in diploids (13,123; 25.5%) and did not converge. By contrast, the pan-genome reached a plateau indicating that the gene family space of European potatoes was greatly covered by the ten genomes analysed here, whereas the diploids had larger pan-genome size22. Data are presented as mean ± s.d., where mean values of gene families and error bars representing standard variations result from n = 2,000 random samplings during pan-genome construction (Supplementary Methods). c, The number of gene families in simulated and actual tetraploid genomes. The random background distribution was generated by repeatedly selecting four random, haploid genomes (of the 40) and calculating the number of gene families in them. Six of the ten genomes had significantly fewer gene families.
