The PIC values of SCoT and BPS markers are smaller than that of the RAPD marker. It might be due to the lower number of polymorphic fragments of the RAPD marker. The MI can be considered a standard measure of efficiency in defining polymorphisms (Varshney et al. 2007). MI of SCoT and BPS (1.03) doubles that of RAPD (0.57). This character shows the SCoT and BPS marker system can be applied to fingerprinting approaches or assessing genetic variation.
The present research exposes that the Rp of SCoT and BPS primers is higher than that of RAPD primers, except for SCoT36. MI and Rp are suggested to be the most popular marker parameters for choosing informative primers. Thus, primer SCoT21 was determined to be the best primer to distinguish the 276 investigated longjack genotypes.
In the present study, the I value of longjack populations was only 0.27 which is relatively low compared to the mean of 0.5 (Culley 2003). He equaled 0.17 which was slightly lower than some other reports for this species (Osman et al. 2003; Rosmaina and Zulfahmi 2013). According to Monfared et al. (2018) genetic characteristics are influenced by both polymorphic and monomorphic fragments. Thus, a low PPL (66.5%) found in four longjack populations in Thua Thien Hue displayed a low level of genetic diversity (Table 4). Some other studies also found similar a percentage of PPL between 57 and 64 ((Hamrick & Godt 1996). The AMOVA showed that the major part of genetic diversity occurred within the investigated longjack populations in Thua Thien Hue, with 87% based on SCoT and BPS markers and 94% based on RAPD markers.
A study by (Hague & Routman 2016) indicated that the high genetic variation within a population is caused by the high level of gene flow and population mixing. Previously, Rosmaina and Zulfahmi (2013), also reported that longjack maintained much more significant genetic changes within populations than between them. Hamrick et al. (1992) stated that the genetic variation of the woody perennial species significantly occurs within populations, in particular in tropical trees. For out-crossing species, the genetic differentiation among populations ranges between 15 and 38% (Bussell 1999). Genetic differentiation among populations (11%) shows that longjack has a comparable low genetic diversity, lower than the findings reported by Susilowati et al. (2012). The main reason might be that this plant is an outcrossing species and has unisexual flowers, in which male flowers have only one sterile pistil and female flowers have several sterile stamens.
The genetic structure is influenced by many factors, including population size, genetic drift, gene flow, seed dispersal, breeding, or natural selection (Hamrick and Godt 1996). The genetic structure of a population can be characterized based on Nei’s genetic distance, AMOVA, and ΦPT. The genetic structure of four longjack populations in Thua Thien Hue exhibited a significant subdivision among them, but genetic differentiation was weak (ΦPT = 0.11). In comparison with the results obtained by Single Nucleotide Polymorphisms (SNPs) (Osman et al. 2003) and RAPD (Rosmaina and Zulfahmi 2013), the genetic differentiation in the present study is lower. It indicates an increased gene flow, possibly due to a large distance of seed and/or pollen dispersal.
Hamrick and Godt (1996) reported that wind or insect-pollinated out-crossing species display a higher genetic variation within populations with a low level of genetic differentiation. Longjack seeds are quite large but can be dispersed over long distances by rainwater flow, birds, or rodents (Susilowati 2008). When seed dispersal is accomplished by these vectors, plants often have high genetic variation within populations (Hamrick et al. 1992).
According to the present results of genetic distance, Nam Dong and A Luoi should be regarded as one group, while Bach Ma and Phong Dien present a second group. These conclusions have nearly a similar range of genetic distance among populations. This result are supported by the PCoA analysis, which also divides these four populations into two groups without a complete clear-cut distinction. The findings of the Bayesian clustering analysis of the genetic structure using STRUCTURE as software presents that the examined longjack populations are best to fit into two potential genetic groups (K = 2), consisting of Bach Ma, Phong Dien (group 1) and Nam Dong, A Luoi (group 2).
The results of the PCoA and STRUCTURE program are generally corresponding to the genetic relationships described in the dendrogram of the Ward’s method. The populations of Nam Dong and A Luoi have comparatively higher levels of genetic diversity in comparison to those of Bach Ma and Phong Dien. As Nam Dong, Bach Ma and A Luoi are geographically closer to each other than to Phong Dien, it is hypothesized that the populations in the mountains should have the same genetic diversity. However, Bach Ma and Phong Dien unexpectedly, create one group separate from Nam Dong and A Luoi.
The reason for these unexpected results may be firstly: the populations are located in the mountains and one in the sandy area. With around 140 km, the populations are not so far away from each other. Although Bach Ma is more adjacent to Nam Dong than to Phong Dien (sandy area), it is close to deltas and sandy regions. Besides, longjack seeds can be transferred by the lagoons and stream river (water flow) along the province through the sandy area districts, which are close to Phong Dien district (sandy area). Several previous studies have reported that the pollination process often occurs in a limited distance (few hundred meters), but Kremer et al. (2012) and Hague and Routman (2016) observed a movement of pollen further than 100-km distance. Therefore, water flow and wild birds might possibly be the reasons why the fruits were transferred from Bach Ma to Phong Dien.
Secondly, the results highlight that the Bach Ma population has the lowest genetic diversity in comparison with the other populations. The number of analyzed accessions in this study may influence the low level of genetic variation of this population as Bach Ma has the lowest quantity of examined individuals (23 DNA samples). According to Egbadzor et al. (2014) and Kapoor et al. (2000) the higher the number of samples is, the more significantly the genetic variation is detected. As Bach Ma has a low number of polymorphic loci and genetic diversity, this may affect the findings of the genetic structure. This is in contradiction to the finding that Phong Dien which is the site with the highest number of analyzed samples, shows lower variation parameters in comparison to Nam Dong and A Luoi. This result suggests that the number of individuals in our study may not have any influence on diversity factors, which is confirmed by other studies e.g., Chen et al. (2017).
Thirdly, during the Vietnamese-American war (1955–1975), Agent Orange weapons seriously destroyed the mountainous areas, especially A Luoi and Bach Ma (Robert, 2016). The United States sprayed herbicides in Vietnam to defoliate 1.1 million ha with ca. 10% of the country (Westing 1986). The war has dramatically influenced the canopy cover, flora and fauna biodiversity and forest genetic resources. Once the forests were destroyed, its fragmented areas also changed the future forest succession. Regarding geographic distance, Nam Dong is close to A Luoi. The pollination (gene flow) may occur from Nam Dong to A Luoi. After the war, A Luoi surely had some new invasive plant species. The former species (e.g., longjack) continue to develop by support from other areas (Nam Dong or Tay Giang area of Quang Nam province). Thus, the war factor and geographic distance are probably one more reason to create the same genetic group containing the sites of A Luoi and Nam Dong.
Finally, the field survey in Bach Ma indicated that it is not easy to find longjack individuals in areas higher than 700 m in elevation. Thus, the geographical and ecological factors might limit the distribution of this species and could probably contribute to the isolation of the populations of Nam Dong and A Luoi. This is confirmed by the Mantel test, which indicates a significant correlation between geographic distances (including elevation factor) and genetic diversity.