Genetic diversity and recombination in P. oryzaelineages from YYT
Analysis of population subdivision in P. oryzae revealed the coexistence in YYT of previously described worldwide lineages along with newly detected lineages specific to YYT. Both the NJ tree (Fig. 2A) and the DAPC barplot (Fig. 2B) estimated from microsatellite data for 513 YYT isolates (215 isolates already analysed in Liao et al. 2016, 298 isolates unique to this study) and 44 worldwide representative isolates, showed that some YYT isolates grouped with previously described worldwide lineages (named W-lineages on the figures) whereas a large group of YYT isolates formed another group specific to this region, subdivided into several clusters. As expected from Liao et al. (2016), isolates coming from glutinous japonica landraces from YYT formed a separate clade, with some spillover genotypes on indicalandraces (Fig. 2B). The coexistence of multiple P. oryzaelineages in YYT was further confirmed by whole-genome analysis of the 46 YYT isolates. Short-reads whole genome re-sequencing of these 46 isolates yielded an average coverage depth of 5X, that resulted in a final dataset of 66,102 SNPs without any missing data after mapping against the 70-15v8 reference genome. These data were pooled together with whole-genome SNPs data from 48 worldwide representative isolates previously sequenced (Gladieux et al., 2018) to build a neighbour-net network using Splitstree. This analysis confirmed that nine isolates from YYT were assigned to two of the four worldwide lineages previously described (Gladieux et al., 2018; Latorre et al., 2020; Thierry et al., 2021) (Fig. 2C: four isolates from YYT assigned to W-Lineage 1, five isolates from YYT assigned to W-Lineage 3), whereas the 37 remaining isolates were assigned to three well-supported lineages specific to YYT, hereafter named lineages YYT1 to 3 (Fig. 2C; Supplementary Information SI2 Table SI2.1). Our analyses thus supported a population subdivision into five genetic lineages. Migration across these lineages was further explored using the TREEMIX analyse, supporting the above structure with possibility of migration between the YYT and worldwide lineages (Supplementary Information SI2 Fig. SI2.3).An overall high genetic variability was observed based on microsatellite data for the entireP. oryzae population sampled in YYT (513 isolates), as shown by gene diversity (0.566), Simpson’s diversity (0.947) and evenness (0.199). At the lineage level (Fig. 2C), nucleotide diversity estimated from full-genome SNPs data was the highest for YYT isolates from worldwide lineage 1 (π=2.00 × 10-4/bp) and was the lowest for isolates from lineage YYT3 (π=0.12 × 10-4/bp ) (Table 1). As compared to π estimates in non-YYT isolates from worldwide lineages provided previously [23], π estimates in YYT isolates assigned to worldwide lineages were either equal (for WL1: π=2.00 × 10-4/bp, this study, vs π=2.11 × 10-4/bp, [23]), or slightly higher (for WL3: π=0.53 × 10-4/bp, this study, vs π=0.45 × 10-4/bp, (Gladieux et al., 2018)). Sequence divergence among lineages was generally higher than nucleotide diversity within lineages, with dxy ranging between 0.09 × 10-3/bp between lineages YYT1 and YYT2 to 2.25 × 10-3/bp between lineages YYT2 and W-Lineage 3 (Table 1).
Signatures of recombination were searched within the P. oryzaepopulation sampled in YYT. For the whole P. oryzae population sampled in YYT (513 isolates), the global linkage disequilibrium (rD) estimated from microsatellite data was 0.145: although significantly different from 0 (P-value=0.001; rD=0 is the expected value under the null hypothesis of free recombination), this value was far from 1 (expected value under complete clonality). The rate of sexual reproduction estimated from microsatellite data among these 513 isolates varied from 15% to 100% depending on the time period considered (Supplementary Information SI2 Table SI2.2). Phi-tests performed within each lineage using whole-genome data were all significant, allowing to reject the null hypothesis of strict clonality, thus supporting the occurrence of recombination. This was supported by the reticulations observed in the minimum networks for lineage YYT1 and, to a lesser extent, for lineages WL1 and WL3 (Supplementary Information SI2, Fig. SI2.2). However, the patterns of LD decay along the genome, estimated within each of the five lineages inferred in YYT, showed limited support for recombination. Finally, we searched the mating type genes in the de novo assembled genomes of the 46 fully sequenced isolates from YYT using BLAST, and confirmed the presence of both mating types in the whole P. oryzae sample. However, only isolates assigned to WL1 comprised both mating types, the other lineages carrying a single mating type (Supplementary Information SI2 Table SI2.3). Altogether, the role of recombination could not be excluded, though a greater number of isolates needs to be genome sequenced to identify its exact role in various lineages.