Table 1.
List of S. marianum accessions used in the present study.
Accession number, DArT sample code, Accession origin: ISO code of the country where the accession was originally collected; Species; Accession description; Donor code: FAO code of donor institutions; and donor accession number were shown.
Fig 1.
Leaf, stem and flower biodiversity of the ex-situ Silybum marianum used in this study.
A) G20 stem. B) G8 striated stem. C) G9 white inflorescence D). G5 non variegated leaf, E) G20 purple inflorescence.
Fig 2.
SNP density plot showing the number of variants within 1 Mb window size along the S. marianum genome.
The horizontal axis shows the chromosome length (Mb); the different colour depicts SNP density.
Fig 3.
Population structure of Silybum accessions using DArTseq technology.
A) Principal Component Analysis (PCA) using high-quality SNP markers. Samples are colored based on their grouping. B) Heat map of kinship matrix created using GAPIT3 [42]. The color histogram indicates the distribution of coefficients of co-ancestry, with the stronger red color showing individuals more related to each other. C) Dendrogram obtained through nonparametric hierarchical clustering. D) bar-plot describing the population Admixture by the Bayesian approach. Each individual is represented by a thin horizontal line, which is partitioned into K-colored segments whose length is proportional to the estimated membership coefficient (q). The population was divided into three (K = 3) groups according to the most informative K value. The colors indicate the accession membership to the groups identified with the Bayesian analysis.
Table 2.
Summary of the analysis of molecular variance (AMOVA) within and among S. marianum populations.
Fig 4.
Phenotypic distribution of oil constituents (A), silymarin components (B) and fruit morphological parameters (C), among the three identified groups (Pop1, Pop2, and Pop3) of S. marianum collection. Boxplots represent the distribution of each trait, with the central line indicating the median, the box edges representing the first and third quartiles, and the whiskers extending to 1.5 times the interquartile range. Outliers are represented by individual black points. p-values are displayed below each boxplot.
Fig 5.
Results of the Bayescan 1.2 outlier test.
Posterior probability significance threshold (vertical bar) of 0.90 after Bonferroni correction (a = 0.05). The locus number ID assigned by Bayescan 1.2 to each marker is reported only for putative outliers SNP.
Table 3.
Frequency alleles of outlier SNPs detected by Bayescan in the three S. marianum groups.
The DArT ID, chromosome, position and candidate genes were also provided. DArTs with allele frequencies (AF) ≤ 0.5 were scored in orange, whereas those with AF ≥ 0.5 were plotted in green.
Fig 6.
Boxplot of DArTs putatively under natural selection with significant effects (p-value < 0.0001) on phenotypic traits of oil constituents in orange, fruit morphological parameters in blue, and silymarin components in green.
For each selected DArT, the germplasm lines were divided into two groups according to their genotypic state (homozygous for reference or alternate allele). The X-axis represents the two alleles for each DArT, while the Y-axis corresponds to the mean of the selected phenotypic trait. Boxplots represent the distribution of each trait, with the central line indicating the median, the box edges representing the first and third quartiles, and the whiskers extending to 1.5 times the interquartile range. Outliers are represented by individual points.