FIGURE LEGENDS
Figure 1. Population subdivision assessed with
FastStructure within the total sample containing 92 YYT and 216
worldwide rice accessions. Each horizontal barplot represents
subdivision assessed for one value of K, the number of inferred
clusters, with K varying from 2 to 10. Vertical lines within each
horizontal barplot represent the probability of ancestry of each
individual in each inferred cluster. Vernacular names of YYT landraces
and names of O. sativa subspecies and of wild Oryzaspecies are indicated below the barplots.
Figure 2. Analysis of subdivision of the P. oryzaepopulation sampled in YYT. A. Phylogenetic relationships based on
microsatellite data of 512 isolates from YYT (green dots) and 43
isolates representative of previously described worldwide lineages
(W-Lineages, non-green dots). B. DAPC clustering analysis based on
microsatellite data of 512 isolates from YYT and 43 isolates
representative of previously described worldwide lineages. C.
Phylogenetic network based on whole-genome SNPs data of 46 isolates from
YYT (red dots) and 48 isolates representative of worldwide lineages
(W-Lineages, non-dotted branches) inferred using the neighbor-net
method. Worldwide isolates used in Fig. 2A, B and C were chosen to
represent the four worldwide lineages defined in previous studies
(Saleh et al. ,
2012; Gladieux et al. , 2018b; Latorre et al. , 2020;
Thierry et al. , 2021).
Figure 3. Comparison of genetic subdivision among 46 rice /P. oryzae paired samples. Left part of the figure: rice
accessions analysed for 26,860 SNPs. Right part of the figure:
corresponding P. oryzae isolates analysed for 66,102 SNPs. The
genealogical trees were built using RAxML, with branch supports after
100 bootstrap replicates indicated on branches. The barplots show the
result of DAPC clustering analysis for a number of genetic clusters (K)
varying from 2 to 10 (each barplot indicates probabilities of ancestry
of each individual in the corresponding K clusters). Black lines in the
central part of the figure connect each rice accession to its
corresponding paired P. oryzae isolate.
Figure 4. Heatmap of quantitative interactions between 33P. oryzae isolates (by lines) and their corresponding rice
accessions (by colomns). The plotted variable is the adjusted
performance (log-transformation of the mean percentage of diseased leaf
surface over 4 scored leaves, estimated from the ANOVA model) and varies
from -4 (low performance) to 0 (high performance). Rice accessions are
coloured according to the genetic cluster they belong to (according to
clustering analysis of rice genotypes, see Fig. 3 left panel). Coloured
dots besides P. oryzae isolates’ names correspond to the genetic
lineage P. oryzae genotypes belong to (according to clustering
analysis of P. oryzae genotypes, see Fig. 3 right panel); (green:
lineage YYT1; red: lineage YYT2; pink: lineage YYT3; yellow: worldwide
lineages 1 and 5; blue: worldwide lineage 3). A: Heatmap without
hierarchical clustering; rice accessions are ordered by rice genetic
clusters, P. oryzae isolates are ordered so that paired samples
(framed) are along the diagonal. B: Heatmap with hierarchical clustering
by lines and columns; branches with bootstrap support above 50 are bold
(10,000 bootstrap replicates); paired samples are framed.
Figure 5. Boxplots of the average performance of P.
oryzae isolates sampled on plants of from a given landrace, measured on
plants of this landrace and of the other landraces. P-values of the
corresponding contrasts are indicated above each set of boxplots.
Table 1. Nucleotide diversity ( within, and sequence
divergence (dxy) between, genetic lineages estimated from whole-genome
SNPs data for 46 P. oryzae isolates from YYT. Both statistics
were estimated in non-overlapping 10kb windows and the average value per
base pair is presented. W-lineages: worldwide lineages as described in
Gladieux et al. (2018). Assignment of the 46 isolates from YYT to
worldwide lineages is based on Fig. 2C & Table SI2.1.