Discussion
The analysis on C. chinensis shows that the contributions of various biotic and abiotic factors to individual physiological states and population structure can be quantified by integrating various information on population with DNA methylation analysis, and then the population dynamics can be narrowly estimated.
During population development from expansion or early stage maturation (in the recovering stand) to the top stage of maturation (in the native stand), accompanied by the change of habitat environment and individual DBH range, relative contribution of soil and spatial factors to methylation decreased, and MVPS variation spatial autocorrelation of low DBH subpopulations disappeared (Fig. 4a–d, Table 1). All of that indicates that, following the development of C. chinensispopulation, the effects of soil and spatial factors on individual development decline and population structure tends to disorder. The spatial epigenetic and genetic autocorrelation represents the aggregation distribution of functional group and kin individuals, respectively (Wang et al ., 2012; Huang et al ., 2015). The disappearance of epigenetic spatial autocorrelation, which accompanied the decline in individual amount but increase in mean DBH (Fig. 1d-e), indicates a functional self-thinning process. This functional self-thinning may result from the competitive exclusion between same functional individuals, then a functional diversity distribution around high DBH individuals forms.
At an individual level, the positive and negative correlations of DBH with the full-methylation rate and hemi-methylation rate are consistent with the result that the relative growth rates decline as the DBH increases in C. chinensis (Fig. 3a,b), because full-methylation and hemi-methylations determine the repressed degree of individual vitality and the potential of development reprogramming. Furthermore, following the increase of edaphic N concentration and decrease of edaphic AP concentration from the recovering stand to the native stand, soil AP constraints aggravate (Fig. 1c) (Hou et al . 2014; Chenet al ., 2016; Turner, Brenes-Arguedas & Condit, 2018). However, AP lost its significant effect on DNA methylation variation in the native stand, in which over 75% individuals’ DBH is over 40 cm and the DBH effect on total full-methylation rate rises to 8.3% through interactions with environmental factors (Fig. 2b–g, Table 1). DBH as an important plant phenotypic indicator may affect individual physiological characteristics (Bustos‐Segura et al ., 2017). The transformation of plant nutritional needs provides a new way for us to redefine nutrient limitations during primary succession of soil formation in subtropical forests (Fig. 2b–g, Table 1) (Turner et al ., 2018). The temporal change on both from individual active to inactive and the transformation of plant nutritional needs are favourable for population maintenance.
The presented approach has several advantages. First, the statistical data of genome-wide, methylation rates that determine individual physiological state and the MVPS that determines the reaction model represent individual characteristics well enough (Suzuki & Bird, 2008). In addition, because the variation of methylation rate and MVPS are direct-acting results of biotic and abiotic factors on individuals and are the direct reason for individual physiological adjustment and population structure change (Schubeler, 2015), all analyses do not create uncertainty because of indirect relationships between variables. Thus, the parsing results and their variation trends of the contribution of various biotic and abiotic impact factors on individual DNA methylation are the closest presentation for a logical and a time relationship of individual development fitness to microhabitat. Similarly, epigenetic spatial autocorrelation analysis results are the intuitive description of population functional group structure.
Second, both the MRM analysis and multivariate variation-partitioning analysis can be employed in non-parametric or non-linear analysis, and their analytical precision can be improved by expanding the number of explanatory matrices, allowing more environmental variables to be represented by their own distance matrices (Goslee & Urban, 2007; Oksanen, 2015). For example, as a stable influencing factor, topography abidingly affects surface runoff and the distribution of soil elements (Latiff, 2009) and also relates to light, humidity and temperature in microhabitats. How topographic factors affect plant physiological characteristics can be illuminated by adding those microenvironmental and microclimate factor matrices into analysis after permutational forward model selection.
Third, DNA methylation contains both instantaneous and cumulative information (Suzuki & Bird, 2008; Heard & Martienssen, 2014; Schubeler, 2015; Harrison et al ., 2016; Heer et al ., 2018; Moler et al ., 2018). DNA methylation parameters in this research were obtained from leaf organisation, which renewed their response mechanism to both the instantaneous environmental condition and age-related DBH indicator. The implications of this bilayer extend the range of methylation research to individuals and populations on time scales.
Many studies on the relationship between genetic structure and epigenetic structure have reached different conclusions (Herrera, Medrano, Mónica & Bazaga, 2016; Heer et al ., 2018; Moleret al ., 2018). For different spatial scales and different species, the contribution of genetic background to DNA methylation may be different. Therefore, that contribution should always be a noted in research. On the other hand, C. chinensis has many common characteristics of constructive species of a subtropical, evergreen, broad-leaved forest, such as being perennial, outcrossing and occupying the top space of the community. All those characteristics may reduce genetic diversity within the population, and then the relationship between genetic structure and epigenetic structure (Herrera et al ., 2016; Moler et al ., 2018). Thus, the functional self-thinning and its effect on population development and community construction, which appears in C. chinensis populations, may provide enlightenment in within-population research of other subtropical constructive species.
The presented approach still has some room for improvement. First, DNA methylation analysis can join with a next-generation sequencing platform (Xia et al ., 2014). The new method can eliminate constraints associated with the methylation-resolving power of the gel and explain adaptive variation at gene level. Secondly, besides the three most distinctive DNA methylation indicators used in this study, other DNA methylation indicators, such as epigenetic distance between individuals, also represent some detailed information on population function diversity (Herrera et al ., 2016). How to integrate more DNA methylation indicators into this presented approach is an interesting topic.