INTRODUCTION
Disentangling the factors as drivers of genomic and phenotypic divergence is essential to understand speciation processes. Allopatric divergence has been considered the most likely cause of speciation for a long time, and targets of natural selection may contribute to allopatric speciation when populations encounter different selective pressures in different habitats (Wiens & Graham, 2005). One of the key speciation forces is evolutionary divergence driven by adaptation to the environment essentially proposed by Charles Darwin (1859). At the genomic level, loci with strong population differentiation reflect local adaptation in populations under different environments, in which reproductive isolation (RI) may appear as a byproduct of the accumulation of genetic differences (Schluter, 2001; Sobel, Chen, Watt, & Schemske, 2010; C. I. Wu, 2001). Natural selection thus acts as a “barrier” to gene flow to produce local genomic divergence between lineages (Arias, Van Belleghem, & McMillan, 2016; Edelman et al., 2019; Tavares et al., 2018). Similarly, gene flow and drift can also act and even interact in several ways during evolutionary divergence (Han et al., 2017; Ma et al., 2018). Geographic patterns of phenotype divergence also reflect that traits have diverged by environmental gradients as selective forces and created adaptive genetic differences (Ayoola et al., 2021; L.-F. Li, Cushman, He, & Li, 2020; T. Zhang et al., 2021). Extensive studies have indicated that morphological divergence is expected to be associated with prezygotic reproductive barriers. Plants can directly or indirectly reduce matting, sperm transfer, or fertilization (Coyne, 2016; Feder et al., 1994) by changing both the flowering time and mating system. For many taxa, includingAquilegia , floral characteristic differences leading to pollinator shifts likely contribute to RI between populations (Des Marais & Rausher, 2010; Hodges, Whittall, Fulton, & Yang, 2002; Kuriya, Hattori, Nagano, & Itino, 2015; Quattrocchio et al., 1999; Schwinn et al., 2006). If populations with divergent selection maintain geographic isolation, environmental differences may drive lineage adaptation and differentiation and play to form RI between them with restricted gene flow. Consequently, to make correct inferences about the driving force of lineage diversification, considering all processes along environmental gradients that shape phenotype and genome is important.
During population divergence, driving forces may issue distinct demographic histories, including divergence time, population size fluctuations, and directionality of gene flow. However, it can be difficult to identify which force is likely to have conduced to the current phylogeographical pattern of a certain species. The identification of highly differentiated regions in the genome may have false positive results due to the influence of the unique demographic histories (Krak et al., 2016). Therefore, it is important to reconstruct the past demography of lineages for plants in which interspecific gene flow has been detected widely. Moreover, East Asia occupies a diversiform climatic and geographical environment and is considered a natural laboratory for adaptive evolution (Ficetola, Mazel, & Thuiller, 2017). Evidence from previous studies in this region has uncovered the demographical histories, genetic diversities, and related influencing factors of many taxa (Areces-Berazain, Hinsinger, & Strijk, 2021; Song et al., 2021; S. Wu, Wang, Wang, Shrestha, & Liu, 2022). While the heterogeneous environment of this area might have contributed to lineage divergence, it is still uncertain which might be the key driver.
Aquilegia viridiflora Pall. (Ranunculaceae), is a dominant, perennial herb that is widely distributed in northern China with obvious variation in phenotypes (Z. Wu, Raven, & Hong, 2001). In previous studies, Andrey S. Erst et al. identified A. viridiflora with purple laminae or lilac-blue petals as the new species A. kamelinii (A. Erst, Shaulo, & Schmakov, 2013) and A. viridiflora with dark purple petals in North China as the new speciesA. hebeica (A. S. Erst et al., 2017). However, the publication of these new species lacks support from molecular data. According to field surveys, the phenotypic variation of A. viridiflora is continuous (Figure S1) and presents substantial challenges in species delimitation. Therefore, A. viridiflora , A. kamelinii and A. hebeica were considered the species complex (A. viridifloracomplex) in our study. In addition, Aquilegia is a well-known example of an evolutionary biology (Kramer, 2009). Phylogenetic studies have defined adaptive radiation in Aquilegia , involving a wide variety of habitats, and divergent selection played an important role in the radiation of Aquilegia (Fior et al., 2013; M. Li et al., 2019; W. Zhang, Wang, Dong, Zhang, & Xiao, 2021). Therefore, theA. viridiflora complex provides a remarkable system for assessing how evolutionary forces in various ways could have shaped genomic divergence patterns during speciation.
In the present study, we collected population samples covering the main distribution range of the A. viridiflora complex for genome resequencing and morphological characteristic measurements. Moreover, we assumed that environmental heterogeneity may put selective pressures on the A. viridiflora complex, driving population divergence and producing genetic variation. Thus, our study focuses on exploring phenotypic and genomic patterns of divergence, and especially on inferring the influence of gene flow, genetic drift, divergence time, divergent selection, and geographic isolation during the speciation process. The investigation reveals the contribution of evolutionary forces to the genomic divergence patterns of the A. viridifloracomplex and helps us understand how demographics contribute to speciation.