PMC

Model for Ug99 origin by somatic hybridisation and nuclear exchange. The ancestral isolate of the Ug99 lineage acquired the A and C genomes from an isolate of the 21 lineage and an unknown isolate and later gained virulence to wheat cultivars carrying the Sr31 resistance gene

Pgt21-0 and Ug99 share one nearly identical haploid genome. a–d Dot plots illustrating sequence alignment of complete haplotypes. X- and y-axes show cumulative size of the haplotype assemblies depicted by coloured bars to the right and top of the graphs. Colour key indicates sequence identity ratios for all dot plots

Gene content of Pgt21-0 chromosome pseudomolecules. a Circos plot showing location of orthologous gene pairs in the A and B chromosomes of Pgt21-0. b Gene and repeat density plots for homologous chromosomes 14 A and 14B. Density of genes encoding non-secreted (black) or secreted proteins (red) along the chromosomes are shown, with individual genes indicated by black or red dots. Bottom graph shows density of repeat elements (blue). Positions of AvrSr50 and AvrSr35 genes are indicated

Somatic hybridisation in Pgt evolution. a Phylogenetic analysis of Pgt isolates from diverse countries of origin (colour key) using a RAxML model and biallelic SNPs called against the full dikaryotic genome of Pgt21-0. Scale bar indicates number of nucleotide substitutions per site. Red asterisks indicate P. graminis f. sp. avenae isolates used as an outgroup. b Dendrogram inferred using biallelic SNPs detected against haplotype A of Pgt21-0. c Dendrogram inferred using SNPs detected against haplotype B of Pgt21-0. d Dendrogram inferred from SNPs detected in haplotype C of Ug99

Strategy to identify homologous contigs in genome assemblies by gene synteny. To detect shared content, Pgt gene models (grey and coloured boxes) were aligned to the genome assemblies and the contig positions of the top two hits of each gene were recorded. Contigs containing at least five shared genes were considered as potential haplotype pairs. Sequence collinearity between contigs was assessed by alignment, and homologous matching contigs were assigned to bins. Examples shown are for Bin04 and Bin12 from Pgt21-0 and Ug99, respectively

Models for the emergence of the founder isolate of the Pgt Ug99 lineage. a A somatic hybridisation event and nuclear exchange occurred between an isolate of the Pgt 21 lineage and an unknown Pgt isolate. The combination of nuclei A and C yielded the parental isolate of the Ug99 lineage in Africa. Under this scenario, nucleus A of Ug99 is entirely derived from nucleus A in Pgt21-0. b Alternatively, sexual reproduction and mating between these two parental isolates defined the origin of the Ug99 lineage. Under this scenario, meiotic recombination and chromosome reassortment would result in the Pgt21-0-derived A nucleus of Ug99 being composed of a mosaic of the two haploid nuclear genomes of Pgt21-0 (X and Y)

Chromosome sets of haplotype A and B in Pgt21-0. a Schematic representation of assembled chromosomes for Pgt21-0 of each haplotype (scale bar = 1 Mbp). Horizontal bars indicate telomeric repeat sequences. b Dot plot of sequence alignment of Pgt21-0 chromosome pseudomolecules of haplotypes A and B. Two translocation events, one between chromosomes 3 and 5 and one between chromosomes 8 and 16, are evident. c Dot plot of sequence alignment between chromosomes from haplotype A in Pgt21-0 and contigs from haplotype A in Ug99. d Percentage of Hi-C read pairs linking each A haplotype chromosome to other A chromosomes (blue) or to B haplotype chromosomes (orange). e Percentage of Hi-C read pairs linking each B haplotype chromosome to either A (blue) or B (orange) chromosomes

A common haplotype containing AvrSr50 and AvrSr35 is shared between Pgt21-0 and Ug99. a Diagram of genomic regions containing AvrSr50 and AvrSr35 alleles in Pgt21-0 and Ug99. Numbers above tracks correspond to contig coordinates and the sense of the DNA strand is indicated as + or −. Predicted gene models (including introns) are depicted as dark grey boxes and intergenic spaces are shown in light grey. Coloured arrows indicate location and direction of AvrSr50 and AvrSr35 genes, with size and position of insertions shown in yellow. Intergenic distances between AvrSr50 and AvrSr35 are indicated by brackets. b Total sequence identity between contigs representing homologous chromosomes of different haplotypes (coloured bars) containing the AvrSr50/AvrSr35 locus (dotted white boxes). Telomere sequences are represented in grey. Chromosome size = ~3.5 Mbp

Haplotype assignment by read subtraction and mapping process. a Illumina reads from Pgt21-0 were mapped to the Ug99 genome assembly at high stringency. Unmapped reads derived from divergent regions of the B haplotype were retained and then mapped to the Pgt21-0 genome assembly. Read coverage of individual contigs with the original and subtracted reads were compared to designate haplotypes as either A or B. b The same process was followed with reads from Ug99 subtracted against the Pgt21-0 reference to designate the A and C haplotypes. c Pie chart showing proportion and total sizes of contigs assigned to haplotypes A, B or C or unassigned in Pgt21-0 and Ug99 assemblies. d BUSCO analysis to assess completeness of haplotype genome assemblies. Bars represent the percentage of total BUSCOs as depicted by the colour key