Pronounced genetic differentiation and recent secondary contact in the mangrove tree Lumnitzera racemosa revealed by population genomic analyses

Abstract

Systematically investigating the impacts of Pleistocene sea-level fluctuations on mangrove plants may provide a better understanding of their demographic history and useful information for their conservation. Therefore, we conducted population genomic analyses of 88 nuclear genes to explore the population dynamics of a mangrove tree Lumnitzera racemosa across the Indo-West Pacific region. Our results revealed pronounced genetic differentiation in this species between the populations from the Indian Ocean and the Pacific Ocean, which may be attributable to the long-term isolation between the western and eastern coasts of the Malay Peninsula during sea-level drops in the Pleistocene glacial periods. The mixing of haplotypes from the two highly divergent groups was identified in a Cambodian population at almost all 88 nuclear genes, suggesting genetic admixture of the two lineages at the boundary region. Similar genetic admixture was also found in other populations from Southeast Asia based on the Bayesian clustering analysis of six nuclear genes, which suggests extensive and recent secondary contact of the two divergent lineages in Southeast Asia. Computer simulations indicated substantial migration from the Indian Ocean towards the South China Sea, which likely results in the genetic admixture in Southeast Asia.

Figure 1. Statistics of genetic variation based on Sanger sequencing data.

(a) Nucleotide diversity π and (b) Nucleotide polymorphism θ at six nuclear genes in 17 populations of Lumnitzera racemosa from the Indo-West Pacific region. Each symbol represents one of the six genes and the average value of six genes from each population. Population abbreviations are defined in Table 1.

Figure 2. Haplotype information of Lumnitzera racemosa in the five populations based on Illumina sequencing data.

(a–c) Show the haplotype networks and distributions of three polymorphic gene segments (Lr191, Lr546b and Lr667a). The haplotype distributions were created by GenGIS version 2.4.0 and MapMaker version 1.0 (http://kiwi.cs.dal.ca/GenGIS/MapMaker). Each colour represents a single haplotype. The circle sizes in the geographic distributions and the haplotype networks are proportional to the sample size of each population and the frequency of each haplotype, respectively. Each branch with more than one mutational step is labelled. Capital letters A and B beside the haplotype networks represent group A and B, respectively. (d) The frequency histogram of group A in 75 highly divergent gene segments from the five populations. Group A is defined based on the criteria described in the Materials and Methods section. Population abbreviations are defined in Table 1.

Figure 3. Haplotype networks and distributions of six nuclear genes across 17 populations of Lumnitzera racemosa in the Indo-West Pacific region.

(a) Lr104, (b) Lr265, (c) Lr559, (d) Lr598, (e) Lr618 and (f) Lr81. The haplotype distributions were created by GenGIS version 2.4.0 and MapMaker version 1.0 (http://kiwi.cs.dal.ca/GenGIS/MapMaker). Each colour represents a single haplotype. The circle sizes in the geographic distributions and the haplotype networks are proportional to the sample size of each population and the frequency of each haplotype, respectively. Each branch with more than one mutational step is labelled. Population abbreviations are defined in Table 1.

Figure 4. Genetic differentiation and population structure across populations of Lumnitzera racemosa in the Indo-West Pacific region.

(a) Results of the BARRIER analysis based on 85 F_ST matrices obtained from Illumina sequencing data. The digital map of this area was created by MapMaker version 1.0 (http://kiwi.cs.dal.ca/GenGIS/MapMaker). The blue line represents a Voronoï tessellation. The red line represents detected barriers and the thickness corresponds to the strength. The number label beside the red line indicates the percentage of genes that support this barrier. (b) A heatmap of the averages of pairwise Kxy values of six nuclear genes among 17 populations. The colour depth is proportional to the level of genetic differentiation. (c) The results of the Bayesian clustering analysis for 295 Lumnitzera racemosa individuals based on six nuclear genes. Each bold bar separated by a black line represents one population and each colour represents one cluster. The best clustering number was estimated to be equal to 2 (K = 2).

Figure 5. Probability density plots of the demographic parameters estimated under the isolation-with-migration model for populations from China (QH), Cambodia (KP) and Malaysia (JH).

(a) A schematic of isolation with migration model. T0 represents the divergence time between the QH and KP populations. T1 represents the divergence time between JH and the most recent common ancestor of QH and KP. (b) Probability density estimation of the divergence times. (c) Probability density estimation of the population sizes for the three populations. (d–f) Probability density estimation of pairwise directional migration rates (2NM). Population abbreviations are defined in Table 1. Note that different scales were used for the horizontal and vertical coordinates in B and E with different colours.