Introduction
As the basic functional and assembly units of communities, populations play a critical role in ecological and evolutionary research. Because individual heterogeneity is ubiquitous in natural populations, to accurately assess population structure, status and destiny requires a basis in an integrative measure of differential responses of different genotypic individuals in different life cycle stages to environmental changes (Valpine et al ., 2014). By now, a robust body of literature has discussed the forming and maintenance mechanisms of plant population (Crone et al ., 2011; Fabritius, Singer, Pennanen, & Snäll, 2019). These papers focus on the genetic analysis of population structure, demographic analysis of population dynamics and physiological analysis of individual adaptive mechanism (Collins & Gardner, 2009; Wang et al ., 2014; Quintana‐Ascencio et al ., 2018). However, different ecological definitions and scales among these three research aspects cut off their mutual connections and hinders us in exploring the spatiotemporal dynamics and developmental mechanisms of population, especially at the within-population level, although ecologists have tended to evaluate the predictive capability of models by integrating more and more population information into meta-matrix models (Crone et al ., 2011; Fabritius et al ., 2019; Plard et al., 2019).
First, although population genetic structure has been extensively studied, neutral molecular markers (e.g., microsatellites or RFLP) used in genetic diversity research have failed to respond to environmental change (Andersen & Lübberstedt, 2003). Variation at marker loci is not an accurate predictor of genetic variation at loci contributing to phenotype, that is, adaptive variation. Because of this, analysis of population genetic structures focuses on seed and pollen dispersal and its effects on gene flow (Browne, Ottewell, Sork & Karubian, 2018). Both seed and pollen dispersal are only preliminary steps for population expansion. The final population colonisation and expansion, as well as the formation of spatial distribution pattern, are determined by the interactions between individuals and habitat selection (HilleRisLambers, Adler, Harpole, Levine & Mayfield, 2012). Second, because of the lack of genetic diversity background of populations, experimental individuals used in physiological analysis can maybe not represent the whole population well enough. At the same time, physiological indicators, such as enzymes, products or substrates of biochemical reaction, among others, only represent one aspect of individual physiological state. Because of the randomness of individual and physiological parameter selection, population research results may be biased (Plard et al ., 2019). Third, in recent years, demographic population models have tried to address individual heterogeneity due to age, size and so on (Plard et al ., 2019). However, those indicators are not correlative with individual basic and instantaneous genetic physiological states. They are just the description of age-related cumulative growth (Browne et al ., 2018, Hamel et al ., 2017; Plard et al ., 2019).
Our research is motivated by those disconnections among research on populations. A key to eliminate the gap among different population research directions is to search for parameters that are definite, measurable and directly related with environment and individual genetic background. Responding to environmental variation, DNA methylation may trigger different gene expression patterns by blocking the promoters at which activating transcription factors could bind and then controlling individual developmental recombination to produce different phenotypes (Suzuki & Bird, 2008; Schubeler 2015). DNA methylation has attracted ecologists’ attention in recent years, but relevant research has surrounded the relationship between population genetic structure and population DNA methylation-variable positions (MVPS) variation pattern and the qualitative relationship between MVPS variation and environmental factors (Heer, Mounger, Boquete, Richards & Opgenoorth, 2018; Moler et al ., 2018). The non-quantitative analysis and DNA methylation parameters’ insufficiency makes it so DNA methylation in population research is still a largely unexplored area.
DNA methylation includes full-methylation and hemi-methylation. Full-methylation was proved to inhibit gene expression in a large body of evidence (Heard & Martienssen, 2014). However, hemi-methylation can either activate or repress gene expression, and activation seems to be more common than repression (Couldrey, Brauning, Henderson, & Mcewan, 2015; Fang et al ., 2016). Hemi-methylated CpG sites were proven to be enriched at core pluripotency loci, and DNA demethylation was enriched in these loci, which promoted somatic cell reprogramming (Heet al ., 2019). Therefore, the DNA hemi-methylation rate is closely related to development reprogramming potential (Liu et al ., 2013; Heard & Martienssen, 2014; Couldrey et al ., 2015; Schubeler, 2015; Fang et al ., 2016; Harrison et al ., 2016; Heer et al ., 2018; Moler et al ., 2018; He et al ., 2019). Genome-wide MVPS variation at is a rounded description of individual physiological reaction (Couldrey et al ., 2015). By appropriately using various epigenetic parameters, DNA methylation can completely represent individual physiological information and then the population spatiotemporal dynamic (Baubec et al ., 2015; Schubeler, 2015).
Here, we use Castanopsis chinensis populations in Dinghu Mountain (DHS) as a model to demonstrate an approach in which we explore the spatiotemporal dynamics and developmental mechanisms of population by integrating various information on population with DNA methylation.C. chinensis is one of the most important constructive species widely distributed from South China to Vietnam. C. chinensispopulations in DHS are in two different development stages. One population has been conserved since it was cut down more than 60 years ago (hereafter, the recovering population). The other population has never been disturbed as it is protected by the nearby Qinyun Temple for over 400 years (hereafter, the native population) (Chen, Rui, Zhou, Ye & Liu, 2016). These populations were studied for their population genetic structures (Wang et al ., 2014), demographic dynamics (Wang et al ., 2014; Chen et al ., 2016), habitat conditions (Chen et al ., 2016) and physiological adaptions. Now, integrating these various studies on C. chinensis population is much-needed for us to narrowly estimate population spatiotemporal dynamics.