Something old, something new: Evolution of Colombian weedy rice (Oryza spp.) through de novo de‐domestication, exotic gene flow, and hybridization

Abstract Weedy rice (Oryza spp.) is a worldwide weed of domesticated rice (O. sativa), considered particularly problematic due to its strong competition with the crop, which leads to reduction in yields and harvest quality. Several studies have established multiple independent origins for weedy rice populations in the United States and various parts of Asia; however, the origins of weedy rice in South America have not been examined in a global context. We evaluated the genetic variation of weedy rice populations in Colombia, as well as the contributions of local wild Oryza species, local cultivated varieties, and exotic Oryza groups to the weed, using polymorphism generated by genotyping by sequencing (GBS). We found no evidence for genomic contributions from local wild Oryza species (O. glumaepatula, O. grandiglumis, O. latifolia, and O. alta) to Colombian weedy rice. Instead, Colombian weedy rice has evolved from local indica cultivars and has also likely been inadvertently imported as an exotic pest from the United States. Additionally, weeds comprising de novo admixture between these distinct weedy populations now represent a large proportion of genomic backgrounds in Colombian weedy rice. Our results underscore the impressive ability of weedy rice to evolve through multiple evolutionary pathways, including in situ de‐domestication, range expansion, and hybridization.


| INTRODUC TI ON
The creation of the agricultural environment through crop domestication provided a new environment for opportunistic plants. Thus in the last 10,000 years, humans have had to contend with the continued evolution of agricultural weeds, which are currently a leading cause of crop losses worldwide (Oerke, 2006). Understanding how agricultural weeds arise and evolve is necessary to devise effective weed management strategies (Vigueira, Olsen, & Caicedo, 2013).
Particularly intriguing from an evolutionary point of view are agricultural weeds that are related to crop species. How plant groups amenable to domestication through artificial selection can also give rise to plants that compete with crops and evade human efforts of removal has long been a question of interest, and may provide insight into mechanisms of plant adaptation (Ellstrand et al., 2010;Vigueira et al., 2013).
The occurrence of weedy rice in multiple and different regions of the world suggests parallel evolutionary events with independent origins. In fact, in recent years, genomic studies have been clarifying the extent to which weedy rice is an example of convergent evolution. Thus, for example, weedy rice from Northeast Asia has been found to have evolved from japonica rice varieties, which are the typical cultivars grown in colder regions in subtropical and temperate zones (Vigueira et al., 2019). Some weedy rice populations in South Asia and in South-East Asia have evolved from indica cultivars, which are typically cultivated in lowland tropical and subtropical regions (Huang et al., 2017;Song, Chuah, Tam, & Olsen, 2014). Other weedy populations from South Asia have been found to have origins from aus cultivated varieties, which are typically grown in Bangladesh, or from local wild rice (O. rufipogon Griff., the ancestor of cultivated rice) (Huang et al., 2017(Huang et al., , 2018. The origin of weedy rice in different regions of the world is thus often influenced by the prevailing local variety and/or the availability of reproductively compatible wild Oryza (Cao et al., 2006;Ishikawa et al., 2005;Olsen et al., 2007).
In the world, Colombia ranks 26th among rice-producing countries and is third in Latin America after Brazil and Peru. Rice is the fourth most important crop in Colombia, after sugarcane, oil palm, and banana. However, rice cultivation in Colombia has reported losses from 30% to 73% due to weeds (Cobb & Reade, 2010). A major weed of cultivated rice in Colombia is weedy rice, for which the presence of 24 plants per m 2 in early-growth stages (40 days after emergence) can reduce crop yields by 50% (Fischer & Ramírez, 1993). Despite its economic importance, no systematic survey of the evolutionary origins of weedy rice has yet been carried out in Colombia.
Several Oryza types could be possible sources of weedy rice in Colombia. The main varieties of cultivated rice grown are routinely released by Fedearroz (National Federation of Rice Growers) and are classified morphologically as indica. A few areas in the western part of the country, collectively known as "Bajo Cauca," are also cultivated with traditional landraces (i.e., "criollos"), whose classification is un- complex.) (Veasey et al., 2011). In Colombia, native wild Oryza species grow mainly in patches of untreated vegetation bordering commercial rice crops; they are found in humid soils, wetlands, and streams, soils with high organic matter content, and can also appear in zones distant from agricultural areas (Villafañe, Estrada, Lentini, Fory, & Palacio, 2007). Oryza latifolia is the species with the greatest distribution, whereas O. grandiglumis has the most restricted range (Estrada, Villafañe, Palacio, Fory, & Lentini, 2010). As with most wild Oryza species, this South American Oryza has traits that are often found in weedy rice, such as red pericarp color, shattering, and seed dormancy.
The plethora of possible contributors to weedy rice in Colombia make the country an interesting case study to contrast with weedy rice evolution in other world regions. Understanding how weedy rice emerges in and adapts to agricultural environments is a necessary first step for designing control strategies for this pest (Vigueira et al., 2013). To determine the origins of weedy rice in Colombia, we have carried out a thorough sampling of weedy rice diversity in the country, as well as sampling of local cultivars and wild species. Here, we use genome-level information to understand the relationships between Colombian weedy rice and local Oryza. We further examine the origins of Colombian weedy rice within an international context, to attempt to understand how the weed evolves and diversifies at a global scale.

| Plant material and DNA extraction
Weedy rice seeds were collected by sampling in the five principal rice production areas of the country in 2014-2015 ( Figure 1).

| GBS library preparation and sequence analysis
Genotyping by sequencing was used to detect polymorphisms distributed throughout the genome among our samples (Elshire et al., 2011).
Restriction digestions were carried out with the enzyme ApeKI, and the fragments were ligated with individual barcoded and common adapters. DNA fragments were pooled for further PCR amplification to enrich the libraries. Single-end fragments of 100 base pairs (bp) were sequenced on an Illumina HiSeq 2500 platform. Raw GBS data have been deposited in the NCBI SRA (experiment SUB6244405).

F I G U R E 1
Map of Colombia showing weedy rice collection areas used in this study. Central zone (red dots), Llanos plains (blue dots), Bajo Cauca river valley (orange dots), North coast (purple dots), and Santanderes area (green dots) Initial data processing was performed at the Biotechnology Institute of Cornell University with the TASSEL-GBS v.3.0 pipeline (Glaubitz et al., 2014) and the MSU7 rice reference genome. The first filters removed SNPs with minimum minor allele frequency <0.01 or missing data per site >90%. Additional filters were applied to the initial VCF file with the NGSEP pipeline (Duitama et al., 2014) at the University of Massachusetts. Final SNPs were supported by at least five reads (for the analysis of Colombian material) or three reads (for the global analysis), and displayed heterozygosity levels of <50%. SNPs with more than 15% missing data and individuals with more than 70% missing data were removed. Only biallelic SNPs were retained, filtering any other type of variants. The final VCF has been deposited in DRYAD.  (Table S1).

| Population analyses
Population structure was analyzed using STRUCTURE (version 2.3.3, Hubisz, Falush, Stephens, & Pritchard, 2009), on the Massachusetts Green High Performance Computing Cluster (http://www.mghpcc. org/). Due to the limitation in the amount of input data that can be handled by the program (Falush, Stephens, & Pritchard, 2003;Pritchard, Stephens, & Donnelly, 2000), approximately 10,000 SNPs were randomly selected for each analysis with a roughly 15,000 base pairs (bp) spacing. Heterozygotes were recorded as "N," and all the simulations were run with data coded as haploid, because weedy and cultivated rice are highly self-pollinated. The program was run for an ancestral "admixture" model, with no correlated allele frequencies.
Runs were carried out using K values between 1 and 15, with three replicates per K, a burn-in period of 100,000 and 500,000 subsequent iterations. The optimal number of genetic groupings was determined using ΔK (Evanno, Regnaut, & Goudet, 2005) according to the program Structure Harvester (Earl & vonHoldt, 2012). The program CLUMPP (Jakobsson & Rosenberg, 2007) was used to obtain a single Q matrix for each K. The final matrix for each K value was visualized with Distruct (Rosenberg, 2004). For comparison, we also analyzed our complete SNP dataset for worldwide samples with the Bayesian clustering analysis fastStructure (version 1.0, Raj, Stephens, & Pritchard, 2014), without recoding heterozygotes, with no prior grouping. FastStructure runs were conducted for K from 2 to 8. The best K was determined through chooseK.py, and the POPHELPER R package was used to generate an image (Francis, 2017).
To investigate the genetic divergence among individuals for all SNPs, the program SmartPCA from the EIGENSOFT package Price et al., 2006) was used. Figures with eigenvalues as coordinates were generated by RStudio 1.0.143.
Genetic diversity for each population was measured by evaluating the expected heterozygosity calculated for all loci, and paired F ST was used to estimate the genetic differences among populations with the software ARLEQUIN (ver 3.5.2.2., Excoffier & Lischer, 2010). Additionally, an AMOVA was performed to analyze variation among and within populations.

| Shattering (sh4) and pericarp color (Rc) gene sequencing
Seventy-one weedy rice accessions from the five rice production areas in Colombia, five commercial rice, and ten landraces, for a total of 96 accessions, were selected for candidate gene analysis. (2010) and DQ421813 reported by Li, Zhou, and Sang (2006), were included. Trees were plotted in iTol v4 (Letunic & Bork, 2016).
Seed shattering and pericarp color were assessed for sequenced accessions. The determination of pericarp color was visually performed at the mature stage of the grain taking into account the intensity of the color. When there were doubts about the color due to the differences in weedy rice maturity in the field, the 2% potassium hydroxide (KOH) test was performed (Louisiana State Seed Testing Laboratory, 1980;Rosta, 1975), immersing dehulled seeds in this solution for 10-20 min. After this time, the KOH solution turns red or dark orange if the seeds are red, while the white seed shows a light golden yellow color or remains colorless.
Shattering measures were taken 30 days after flowering using a digital force gauge to measure breaking tensile strength (BTS), and averaging values for 10 seeds in the same panicle . Statistical analysis was performed by analysis of variance (ANOVA) of one factor, using the statistical package SPSS version 23. The comparison of means was performed by the Tukey test at p < .05.  Table S4).

| Colombian weedy rice origins
Both population structure and PCA suggest no contribution of native wild Oryza to CWR; thus, we excluded these groups and re-analyzed population structure with the available 10,923 SNPs. Three The ancestral contributions in CWR could stem directly from local (Colombian) cultivars that have undergone de-domestication, or from dispersal of weedy rice from other world regions into Colombia. We constructed a phylogenetic tree with our global set of samples, including out-groups, and note CWR falling in different portions of the tree ( Figure 5 and Figure S2). In particular, CWR classified as indica-like based on our structure analysis is related to world weedy rice with known indica origins (e.g., US SH weeds, South Asian weeds) and to known indica cultivars. However, the closest relationship is with the Colombian commercial rice varieties, suggesting that local de-domestication events gave rise to indica-like weedy rice in Colombia. Of the five accessions of CWR identified as >80% aus-like based on our population structure analysis, two are nested within the BHA clade, which is a monophyletic group of US weeds with known aus origin ( Figure 5). Two others cluster close by, but with low bootstrap support, and one clusters with indica-like weeds from South Asia, also with low bootstrap support. This suggests that aus-like weeds in Colombia are more likely to have originated directly from US weeds and perhaps were inadvertently imported, although an importation event from Asia cannot be completely dis-

| Genetic diversity in Colombian weedy rice
We examined levels of genetic diversity in CWR compared to other Oryza groups. Colombian weedy rice as a whole has lower levels of expected heterozygosity than any cultivar group examined (Table 1).
Among CWR groups, aus-like weeds present a greater genetic diversity, despite their low numbers, than indica-like or weeds of admixed origins (Table 1). When genetic diversity was analyzed for the different rice-growing regions in Colombia, we did not see a substantial difference among regions (Table 1).
We assessed levels of genetic differentiation between CWR groups and major cultivated rice groups, which confirmed the F I G U R E 3 Population structure of 168 Colombian Oryza accessions (140 weedy rice, 9 commercial rice, and 19 landraces). Each individual is represented by a color bar, with color partitions that reflect the relative proportion of genetic membership in a given group (K = 2, K = 5, and K = 7) very low F ST values for indica-like weeds and commercial rice from Colombia (F ST = 0.093), which is of indica type, followed by an also modest F ST for indica-like CWR and indica (F ST = 0.162) (Table S10).
Additionally, the lowest F ST values for Colombian aus-like weeds and crop groups were with aus cultivars (F ST = 0.232). Comparisons of CWR with weedy rice groups distributed worldwide revealed that the lowest levels of differentiation for CWR aus-like weeds were with US weeds of aus origin (BHA; F ST = 0.14; Table S11). This F ST value was lower than that between aus-like weeds and aus, supporting a direct origin from US weeds. Interestingly, the lowest F ST values for admixed CWR were in comparison with indica-like CWR, supporting the greater genetic contribution of this group to these admixed weeds.  Figure S3). The fixation of the derived T allele occurs despite the primarily shattering phenotype displayed by CWR ( Figure S3, Table S13).

F I G U R E 4 Population structure of 548
Colombian and global Oryza accessions. Each individual is represented by a colored bar, with color partitions that reflect the relative proportion of genetic membership in a given group. (a) Results of the analysis of Colombian accessions (140 weedy rice, 9 commercial rice, and 19 landraces), Asian (128 cultivars, 173 weedy rice, and 53 wild rice), and United States (9 cultivars and 17 weedy rice) at K = 2, K = 3, and K = 4 populations. (b) Population structure of Colombian weedy rice according to the rice geographic areas at K = 4 Outer ring colors represent cultivar groups of O. sativa. Colored triangles at the tip of each branch represent the geographic origin for world weedy rice accession. Arrows show accessions of CWR identified as >80% aus-like. Individual accession IDs can be found in Figure S1 possessing red pericarps, whereas indica and admixed weeds have some samples with white pericarps (33% and 11%, respectively) ( Figure 7b).

| Pericarp color and the Rc gene
We sequenced 421 bp spanning the seventh exon of the Rc locus for phenotyped accessions. As expected, all accessions with white pericarp (both weeds and cultivated rice) possessed the 14-bp deletion, which has been shown to be responsible for the loss of color in most cultivated rice (Furukawa et al., 2007;Sweeney, Thomson, Pfeil, & McCouch, 2006; Figure S4). For accessions with red pericarps, wild diploid samples of O. glumaepatula lacked the deletion, as expected, as did most of the red pericarp CWR, independent of their ancestry. However, seven weedy rice accession (mainly of indica origin) also with red pericarps carried the 14 bp deletion as well. We examined for the presence of the two additional mutations which have been reported to reverse the effect of the 14 bp deletion on Rc function, re-establishing the red colored pericarp (Brooks, Yan, Jackson, & Deren, 2008;Lee, Lupotto, & Powell, 2009), but neither were found in the accessions evaluated. Our results suggest that in a subset of weeds, other loci are having an effect on pericarp color.

| Colombian weedy rice has multiple origins
Understanding the origin and evolution, genetic diversity, and structure of weedy rice populations is basic knowledge that is necessary for the design of effective methods of weed control (He, Kim, & Park, 2016;He et al., 2014;Vigueira et al., 2013), as well as preventing weedy rice adaptation to crop fields (Huang et al., 2017).
Three main hypotheses have at various times been put forward for the evolution and origins of weedy rice: (a) direct colonization and adaptation of wild rice groups to the cultivated environment; (b) "dedomestication" of cultivated rice varieties that become weedy; and (3) gene flow between cultivated rice and reproductively compatible wild populations, leading to weedy rice of hybrid origins (Vigueira et al., 2013). In studies of weedy rice around the world, examples of all of these origins have been found (e.g., Huang et al., 2017;Li, Li, Jia, Caicedo, & Olsen, 2017;Reagon et al., 2010;Song et al., 2014;Vigueira et al., 2019), underscoring the ease with which weedy rice can repeatedly arise in diverse cultivated environments. Oryza of the CCDD type and cultivated rice with an AA genome likely creates a genetic barrier that reduces the possibility of gene flow between most South American wild species and cultivated rice (Oka & Chang, 1961). The remaining wild species, O. glumaepatula, has the same AA genome as cultivated rice, and there are reports of it hybridizing with O. sativa (Espinoza & Lentini, 2005). However, there is no evidence for hybridization in our study, and we suggest that in several rice-growing regions of Colombia, O. glumaepatula is too geographically limited to have an effect. Although this species grows in Colombia in non-intervention areas bordering commercial rice crops (Estrada et al., 2010), it has a low potential distribution in rice-growing regions of the Caribbean and inter-Andean valleys (Villafañe et al., 2007). Early studies with microsatellites in Colombia reported that some CWR accessions from Tolima (the Central zone) with black hulls and awns showed high genetic similarity with O. rufipogon (González et al., 2003(González et al., , 2007; given the lack of evidence for contribution of this Asian species to CWR, we suggest that this was an erroneous conclusion likely due to the lack of inclusion of the diversity present in rice cultivars, in particular exclusion of aus varieties. Multiple studies have supported de-domestication as the prin- is an evolutionary process that involves a loss of some traits added under domestication, and likely begins with the formation of a "volunteer weed," a crop that acquires a trait by mutation or hybridization, which provides an advantage in the environment (Ellstrand et al., 2010;Gressel, 2005).
In Colombia, two types of cultivated rice are planted, either of which could potentially give rise to weedy rice through de-domes- Commercial cultivars in Colombia have long been presumed to be indica, which is confirmed by our results. These cultivars also seem to be the largest source of weedy rice in Colombia, as most of our sampled CWR was found to be indica-like, clustered close to Colombian cultivars in our tree, and harbored very little genetic differentiation with Colombian commercial cultivars based on F ST .
Our results support growing evidence that in many parts of the world, the origins of this weed are often related to local crop varieties. However, research on weedy rice in the United States has also shown that de-domesticated weeds do not necessarily always evolve from local cultivars, as SH and BHA weedy rice descend from indica and aus cultivars, respectively, neither of which are cultivated  (Gealy, Agrama, & Eizenga, 2009;Londo & Schaal, 2007;Reagon et al., 2010). Such exotic origins of de-domesticated weeds are also evident in Colombia, which harbors a small proportion of weedy rice that is aus-like in genomic back- ground. Yet, there are no records of aus cultivars ever having been grown in Colombia, and its cultivation is restricted to the center and Northeast of India and Bangladesh (Garris, Tai, Coburn, Kresovich, & Mccouch, 2005;Londo, Chiang, Hung, Chiang, & Schaal, 2006), which are the centers of origin of this cultivar (Civáň, Craig, Cox, & Brown, 2015).  (Delouche et al., 2007;Reagon et al., 2010). Similarly, it has been previously suggested that weedy rice in Latin America and the Caribbean was introduced as a contaminant in rice seeds imported from the United States (Hernandez, 1971;Ortiz-Dominguez, 1999). While this is an unlikely explanation for indica-like CWR, it is the most likely explanation for the occurrence of aus-like CWR.
The last, and second most common, group of CWR seems to be 58.8% of the planted area in Colombia was sown with weed-free-certified seed (DANE & FEDEARROZ, 2017). This implies that the remaining areas are being planted with seed that may not be weed free and that often stems from farmers exchanging seed saved from their fields, providing a route for weedy rice to expand its range.

| The source of weedy traits in Colombian weedy rice
As for all de-domesticated weeds, a major question about weedy rice is how it has acquired the traits that enable weediness, which are not typical of their crop ancestors. Two particularly characteristic traits of weedy rice are the dispersal of seeds through shattering and the red pericarp color, which is associated with seed dormancy.
As expected, Colombian weeds showed high levels of seed shatter-  (Huang et al., 2017).
All CWR groups carry the domesticated allele of the gene sh4, which contains the non-shattering T substitution that was fixed during rice domestication. Sh4 is involved in the formation and function of the abscission zone from which a mature grain separates from the pedicel (Li et al., 2006). While there is evidence that cultivated alleles of sh4 are not always sufficient for loss of shattering and are likely dependent on genetic background (Zhu, Ellstrand, & Lu, 2012), sh4 remains among the most significant genes to have experienced selection by humans during domestication (Li et al., 2006;Wang et al., 2018;Zhang et al., 2009 (Huang et al., 2018;Zhu et al., 2012), and Italy (Grimm, 2014). Thus, the results for CWR add to mounting evidence that de-domesticated weeds evolve the shattering trait through novel genetic mechanisms rather than reversals to ancestral alleles. Preliminary mapping studies in SH and BHA weedy rice groups suggest that these novel loci may not necessarily be shared among independently arisen weedy rice groups (Qi et al., 2015;Thurber, Jia, Jia, & Caicedo, 2013). Given that a subset of CWR shares ancestry with BHA weeds from the United States, however, it is possible that the same loci are involved in reacquisition of shattering in BHA as in Colombian aus-like and indica-aus admixed weeds groups.
The proanthocyanidins responsible for pericarp are believed to play an important role in seed dormancy and in the protection against bacteria and fungi in the soil (Sweeney et al., 2007). In weedy rice, seed dormancy allows the seeds to remain in the soil for long periods of time, guaranteeing the permanence of the group (Grimm, 2014). As is common in weedy rice groups (Ziska et al., 2015), most of Colombian weedy rice accessions (77%) have reddish pericarp coloration. These reddish colorations in the pericarp are present in all types of Colombian weedy rice (aus, indica and admixed). In contrast, weeds with white pericarps, which are a minority, only occur in indica-like or admixed weeds, supporting the perceived advantage of red pericarps in weedy rice.
As expected based on prior studies of Rc (Furukawa et al., 2007;Sweeney et al., 2007), the red pericarp in most CWR can be explained by the presence of the functional ancestral Rc alleles that lack the 14-bp deletion that knocks out the gene and leads to white pericarps. Rc is a pleiotropic gene that affects both pericarp coloration and dormancy through promotion of expression of genes involved in the biosynthesis of abscisic acid, which is a dormancy-inducing hormone, and through activation of a conserved network of eight genes involved in synthesis of flavonoids, which are pigments (Gu et al., 2011). The occurrence of wild functional Rc alleles in what is a de-domesticated weed is most likely explained through direct inheritance of this low-frequency allele from domesticated ancestors. Although most cultivated rice has a white pericarp and the 14-bp deletion in Rc, cultivars with red pericarps and ancestral alleles also exist, and this standing variation in crop ancestors could have been selectively favored in weedy rice.
Although no red pericarp cultivars occur in Colombian commercial rice, we cannot be absolutely sure these have not been used for breeding purpose in the past.
A startling find is that a small subset of weedy rice display red pericarps despite having the 14-bp Rc deletion. Restorer mutations in this gene have previously been reported in some cultivars, such as Rc-r (Lee et al., 2009) and Rc-g (Brooks et al., 2008), but we did not find these. Our data are insufficient to explain whether undiscovered restorer mutations have occurred or if other loci are involved in this red coloration, but this finding again underscores the selective advantage that red pericarps must confer on weedy rice.

| CON CLUS IONS
Since the beginning of agriculture, weedy plants have evolved to flourish in crop fields with detrimental effects on agricultural production.
Understanding the source of such plants and how they adapt to agricultural conditions are important pre-requisites for ameliorating their impact crop yields. As studies of weedy rice, a major weed of cultivated rice, accumulate, it has become evident that this weed has had multiple independent origins, most often from domesticated ancestors. Our results from Colombian weedy rice support this growing evidence that the origins of weedy rice are influenced more by domesticated groups than by wild rice, even when local wild Oryza species exist, as several do in South America. De-domestication has additionally been increasingly recognized as a significant mechanism for agricultural weed origins beyond the weedy rice system (Ellstrand et al., 2010), as evidenced by recent studies showing crop origins for Tibetan semi-wild wheat and weedy rye (Burger & Ellstrand, 2014;Burger, Holt, & Ellstrand, 2007;Jiang et al., 2019).
An important consequence of this study is the highlighting of the need to consider weedy rice evolution in any locality within a global context. Although indica-like weedy rice in Colombia most likely has origins in de-domestication from the local indica crop, our data suggest that aus-like weeds are likely exotic introductions, a finding that would have been impossible to discern if we had not compared our Colombian weeds with the range of worldwide diversity in weedy and cultivated Oryza. Our study also underscores how dynamic the evolution of weedy rice can be: Given a local weedy rice origin and an exotic introduction, a new category of admixed weedy rice has successfully emerged in the country. We propose that continued use of a global perspective in weedy rice evolution studies, as well as in other weed evolution studies, will enhance our understanding of the facility with which agricultural weeds can evolve and the mechanisms they use to adapt to the agricultural environment.

ACK N OWLED G EM ENTS
This project was funded, in part, by a Developing Country Collaboration supplement to U.S. National Science Foundation (NSF) grant IOS-1032023 and the Universidad Nacional de Colombia, Bogotá, Special thanks to Fedearroz staff for their help in obtaining cultivated rice accessions and Sherin Perera for help with genotyping. We also thank Jacob Barnett for help with scripts.

CO N FLI C T O F I NTE R E S T
No conflicts of interest have been declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
Raw genotyping-by-sequencing data have been deposited at the NCBI Short Read Archive (experiment BioProjectIDPRJNA616397).