1 Introduction
The evolution of crop species may extend beyond the control of a breeder. Perhaps the most dynamic opportunities for spontaneous evolution are in regions rich with reproductively compatible plants in both managed and unmanaged populations. For example, managed populations may consist of both landraces and inbred rice varieties, while unmanaged populations may include both wild rice and weedy rice, descended from the crop or wild species from which the crop has evolved. According to Molina et al. (2011), the rice crop has been domesticated from wild Oryza species, specifically O. rufipogon Griff. in Asia and O. barthii A. Shiv. in Africa. It is ecologically logical for the epicenters of rice cultivation to have weedy types that are admixtures of the wild relatives, off-types of the selected and cultivated strains, and various outcrosses (Pusadee et al., 2013; Veltman et al., 2019) as they co-exist in adjoining habitats. In such regions, gene flow between cultivated plants, wild relatives, and their weedy descendants present would generate a variation on which artificial and natural selections could act. Such ”domesticated-weed-wild complexes” (hereafter, DWWC) are poised to be genetically rich meta-populations of evolutionary experimentation. While it is easy to tentatively identify such potential complexes by the occurrence of visually apparent hybrids, it is much more difficult to measure whether the gene flow plays a role in the evolution of co-occurring reproductive-compatible relatives (Pusadee et al.,2013; Song et al., 2014). Therefore, the apparent hybrids might be aberrant individuals, or if they are true hybrids, they might be sterile evolutionary dead ends. Genetic analysis is necessary to document whether and how gene exchange occurs (Ellstrand, 2003; Linderet al ., 1998; Leak-Garcia et al ., 2012). Using a panel of 638 domesticated and wild rice genomes, Wang et al . (2017) examined the genetic ancestry of wild rice and demonstrated that most modern wild rice types are heavily admixed with domesticated rice due to gene flow from both pollen and seeds. This suggests that what has been thought of as wild rice may actually be different stages of feralized domesticated rice. Further, the Southeast Asian wild-like weedy rice may have originated from weedy descendants of coexisting cultivars (Sudoet al ., 2021; Neik et al ., 2019).
Most genetic studies of DWWC have focused on one-way introgression of crop alleles into wild or weedy populations (Ellstrand, 2003), whereas only a few studies have examined multi-way gene flow among all members of the DWWC to evaluate whether it is an integrated ”complex”. The island nation of Sri Lanka is an ideal system to study the multilateral gene flow of a DWWC due to its geographical isolation from the rest of South Asia, and its highly diverse topography and climate. The DWWC on the island involves six evolutionarily distinct groups in theOryza genus, including landraces, inbred rice varieties, two wild species, feral populations, and weedy rice. Landraces (traditional rice accessions) have grown for more than two and a half millennia across the island (Weerakoon et al., 2011) and inbred rice varieties cultivated since the 1960s (Dhanapala, 2020). Reproductively compatible free-living wild relatives of the rice include two wild species, the annual O. nivara and the perennial O. rufipogon . Self-sustaining feral populations that have emerged from cultivated fields abandoned (approximately ten years prior) represent an additional category of free-living lineages (Qiu et al., 2020). Nevertheless, their inclusion as part of the DWWC has not yet been recognized. The latter thus presents an unusual opportunity as it represents the early portion of the evolutionary trajectory to de-domestication (Ellstrand et al., 2010). The sixth category is one of the world’s worst rice weeds, weedy rice (Oryza spp.). Unlike feral populations, which are sustained without human intervention, weedy rice is typically restricted to cultivated rice fields, making it a significant threat to sustainable rice production in Sri Lanka (Ratnasekera, 2015). Given the six categories, there are thirty potential pathways for gene flow among the Sri LankanOryza types if all of them participate in the DWWC.
Therefore, our study aimed to use SSR loci to characterize the populations of the entire array of Sri Lankan Oryza categories and address the following research questions: (1) What is the influence of the genetic background of DWWC components and their participation in an integrated complex on the population? (2) To what extent does multi-way gene flow exist among all members of the DWWC? (3) What role has the DWWC played in the origin and evolution of weedy rice in Sri Lanka?