Genetic diversity of 22 T. sinense populations
102 alleles were detected by the 14 SSR loci across the 193 individuals from 22 populations, Original gel of Primer 61 is presented in Supplementary Fig S1. At the species level, the value of Na, Ne, I, Ho, and He were 7.429, 4.435, 1.631, 0.442, and 0.559, respectively (Table S1). At the population level, the value of Na, Ne, I, Ho, and He were 3.221, 2.505, 0.937, 0.434, and 0.566, respectively. The genetic diversity of DFD, TJH, FTZ, FJM, LGM, BDGM, SNJ, WFHH, BSJ and BMXM populations are positive than 1 with high level. The Fixation Index (Fis) value was − 0.082 (Table 2). Ewens–Watterson neutral test results showed that most polymorphic sites did not deviate significantly, which is essentially consistent with the neutral evolution model (Table S2).
Table 2
Genetic diversity of 22 T. sinense populations.
Population | Na | Ne | I | Ho | He | Fis |
LX–HK | 3.214 | 2.313 | 0.939 | 0.481 | 0.519 | 0.077 |
EMM | 2.714 | 2.562 | 0.911 | 0.381 | 0.619 | − 0.131 |
DFD | 3.643 | 2.606 | 1.008 | 0.416 | 0.584 | − 0.102 |
MCM | 2.857 | 2.352 | 0.863 | 0.513 | 0.487 | 0.005 |
TJH | 3.714 | 2.649 | 1.041 | 0.528 | 0.473 | 0.157 |
FTZ | 3.500 | 2.851 | 1.059 | 0.415 | 0.585 | 0.006 |
JFM | 2.714 | 2.397 | 0.816 | 0.368 | 0.632 | − 0.367 |
FJM | 3.500 | 2.740 | 1.033 | 0.453 | 0.547 | 0.021 |
KKS | 3.071 | 2.284 | 0.836 | 0.483 | 0.517 | − 0.084 |
LGM | 3.929 | 2.757 | 1.080 | 0.519 | 0.482 | 0.133 |
DSH | 2.571 | 2.19 | 0.796 | 0.304 | 0.696 | − 0.397 |
SHM | 2.857 | 2.254 | 0.791 | 0.456 | 0.544 | − 0.229 |
BDGM | 3.429 | 2.910 | 1.038 | 0.332 | 0.668 | − 0.187 |
QZMM | 3.071 | 2.600 | 0.96 | 0.253 | 0.747 | − 0.362 |
SNJ | 3.857 | 2.580 | 1.063 | 0.442 | 0.558 | − 0.012 |
WFHH | 3.643 | 2.744 | 1.079 | 0.250 | 0.750 | − 0.261 |
FP | 2.846 | 2.146 | 0.831 | 0.525 | 0.475 | 0.095 |
HBY | 2.231 | 1.952 | 0.657 | 0.444 | 0.556 | − 0.214 |
BSJ | 3.929 | 2.939 | 1.149 | 0.542 | 0.458 | 0.242 |
KP | 2.857 | 2.149 | 0.797 | 0.495 | 0.505 | − 0.088 |
BMXM | 4.000 | 2.944 | 1.121 | 0.417 | 0.583 | 0.017 |
GLGM | 2.714 | 2.189 | 0.737 | 0.540 | 0.460 | − 0.125 |
Mean | 3.221 | 2.505 | 0.937 | 0.434 | 0.566 | − 0.082 |
Note: Na, Number of alleles; Ne, Number of effective alleles; I, Shannon's information index; Ho,Observation of Heterozygosity; He, Expected Heterozygosity; h, Nei's Diversity Index; Fis, wright’s fixation Index. |
Genetic variation and genetic differentiation of 22 T. sinense populations
AMOVA analysis showed that the genetic variation within populations accounted for 84% of the total variance, while the genetic variation among populations accounted for only 16% of the total variance. This indicates that the most genetic variation of T. sinense came from within populations (Table 3). The genetic differentiation coefficient (Fst) at the species level was 0.31, indicating a significant genetic differentiation among populations (Weir and Cookerham 1984). Furthermore, the Nm at the species level was 0.555, indicating that genetic drift was the main cause of genetic differentiation (Table S3).
Table 3
The AMOVA analysis of T. sinense.
Source of variation | DF | SS | Variant components (%) |
Among Populations | 21 | 236.483 | 16 |
Within Populations | 171 | 738.672 | 84 |
All | 192 | 975.155 | 100 |
| Fst = 0.16 | |
Notes: DF, degree of freedom; SS, Sum of Squares of Deviations. |
Genetic structure of 22 T. sinense populations
STRUCTURE analysis showed that there was an obvious inflection point and the maximum value was obtained when K = 3. This indicates that the T. sinense populations could be divided into three clusters (Fig. 3a, Fig. 3b). Cluster 1 included the 9 populations of LX–HK, DFD, MCM, TJH, FTZ, JFM, LGM, DSH and SHM, cluster 2 was composed of the 9 populations of EMM, FJM, KKS, BDGM, QZMM, SNJ, WFHH, BMXM and GLGM, and cluster 3 was composed of the 4 populations of FP, HBY, BSJ and KP (Fig. 3c).
Genetic bottleneck of 22 T. sinense populations
When SMM was used, only SNJ (Na = 3.86, Ne = 2.58, Ho = 0.44, He = 0.56) showed a significant excess of heterozygosity. When TPM was used, only FTZ (Na = 3.50, Ne = 2.85, Ho = 0.42, He = 0.59) had a significant excess of heterozygosity as estimated with the two methods (P < 0.05), suggesting that SNJ and FTZ deviated from mutation-drift equilibrium and indicate that there was a past reduction of effective population size in the species (Table 2, Table 4). Populations of T. sinense underwent a demographic bottleneck in history.
Table 4
Models of Bottleneck Detection of T. sinense.
Population | SMM | TPM |
Sign test | Wilcoxon test | Sign test | Wilcoxon test |
BSJ | 0.131 | 0.735 | 0.534 | 0.497 |
BMXM | 0.410 | 0.761 | 0.388 | 0.194 |
DFD | 0.166 | 1.000 | 0.220 | 0.217 |
FJM | 0.558 | 0.946 | 0.545 | 0.455 |
FP | 0.598 | 0.791 | 0.562 | 0.380 |
FTZ | 0.094 | 0.058 | 0.028* | 0.009* |
KKS | 0.594 | 0.946 | 0.535 | 0.735 |
KP | 0.603 | 1.000 | 0.425 | 0.588 |
LGM | 0.523 | 0.903 | 0.422 | 0.268 |
LX–HK | 0.456 | 0.502 | 0.207 | 0.194 |
SHM | 0.421 | 1.000 | 0.424 | 0.465 |
SNJ | 0.007* | 0.042* | 0.196 | 0.268 |
TJH | 0.465 | 1.000 | 0.562 | 0.244 |
WFHH | 0.315 | 0.839 | 0.572 | 0.414 |
Notes: SMM, Step mutation model; TPM, Two phase mutation model; *: P < 0.05. |
Notes: a, the trends of Ln P (D)(± SD) against K; b, the trends of ΔK against K; c, The clustering results of STRUCTURE(K = 3). |
Demographic history of 22 T. sinense populations
Based on the generated genetic structure, three types of evolutionary models were established, including branching differentiation model, radial differentiation model and common differentiation model (Fig. 2, Fig. 4). Using DIYABC software to speculate on the differentiation time of different clusters, scenario 4 was determined to be the most suitable for the evolutionary model of the nuclear gene of T. sinense as the three clusters have common ancestral population. Each group presents a pattern during evolution. The principal component analysis (PCA) of the simulated dynamic model showed that scenario 4 accounts for 30.9% (Fig. 4a, 4b, 4c), indicating the results are reliable. A posteriori probability results of direct estimation were the same as those calculated through logistic regression (Fig. 5a, 5b), which supports scenario 4 as the best evolutionary model. Supposing the generation time of T. sinense was 50 to 100 years for a generation, the time of differentiation is 1.115×104 to 2.23×104 years ago (ta = 223), the size of the ancestral group is 2.73×104 (NA) and the total size of the progeny population is 1.043×104 (N1 + N2 + N3). The results showed that glaciers and cold climate have become the main factors influencing the existing genetic structure of the T. sinense population with the deterioration of the climate of Quaternary glaciation.