Note: K, the number of clusters; RepTimes, repeat times of independent simulations run for each value of K; lnP (D), the log likelihood of the observed genotype distribution in K clusters that can be output by STRUCTURE simulation (Pritchard et al., 2000); ΔK, an ad hoc quantity related to the second order rate of change of the log probability of data with respect to K, this is a good predictor of the real value of K (Evanno et al., 2005).
Table S7 ITS genotypes (H) and CHS genotypes (C) among the 28 populations
P.

Location

ITS (H)

CHS (C)

1

Huzhu, QH

H1,H2

C1,C2,C3

2

Ledu, QH

H1,H2

C4,C5,C6

3

Xining,QH

H1,H2

C7,C8,C9

4

Haiyan,QH

H3;H3,H4

C10,C11

5

Lintao,GS

H5

C12,C13;C13,C14

6

Luqu,GS

H6

C15,C16

7

Tongren,QH

H7

C16,C17,C18

8

Guoluo,QH

H7

C10,C19,C20;C19,C21,C22

9

Hongyuan,SC

H8

C23,C24

10

Shiqu,SC

H9; H10,H11

C10,C25;C25,C26;C20,C27,C28

11

Jiangda,XZ

H12,H13;H14,H15

C29;C30,C31

12

Yushu,QH

H16,H17

C32

13

Lixian,SC

H18,H19

C33,C34,C35,C36,C37

14

Aerjinshan,GS

H20

C38,C39,C40,C41,C42

15

Dangjinshan,GS

H20

C38,C43,C44,C45

16

Yumen,GS

H20

C41,C46,C47,C48,C49

17

Delingha,QH

H3,H21

C20,C49,C50,C51,C52,C53

18

Xinghai,QH

H22,H23

C46,C54,C55,C56,C57

19

Shandan,GS

H24,H25

C40,C58,C59,C60,C61

20

Zeku,QH

H26,H27

C16,C26,C46,C62,C63

21

Yushu,QH

H28,H29

C20,C50,C64

22

Yushu,QH

H7,H30

C16,C64,C65,C66

23

Yushu,QH

H31,H32

C66,C67,C68,C69

24

Dingqing,XZ

H31,H33

C13,C20,C36,C70

25

Leiwuqi,XZ

H28,H31

C20,C36,C71

26

Ranwu,XZ

H12,H13

C20,C36,C71,C72,C73

27

Dangxiong,XZ

H28,H31

C36,C69,C74

28

Kangma,XZ

H31,H34

C36,C69,C75

Fig. S1 Principal Coordinate Analysis (PCoA) plot based on a matrix of Jaccard distances among 157 individuals using NTSYSPC 2.0. All the individuals could be identified within seven groups, which correspond to the output of the Structure analysis when K=6: the tetraploid populations clustered into three groups, STP (populations in southern QTP), ETP (populations in eastern QTP) and NTP (populations in northern QTP); the diploid populations readily clustered into five groups and DP2 was intermixed with NTP.
Fig. S2 ITS phylogenetic relationships of sampled Allium przewalskianum in the present study using Bayesian, ML and MP analyses. All analyses were performed according to the methods outlined in the main paper. We used Allium ovalifolium, Allium eduardii and Allium mongolicum as outgroups. All newly obtained ITS sequences were deposited in GenBank (accessions KC189069–KC189120). We recovered 34 genotypes from 28 populations (Table S7). The ML tree and the MP tree were congruent with the 50% major consensus tree obtained from Bayesian analysis (shown in this Fig., under the GTR+I+G model) and had a similar topology. The support values are given above the branches (Left: posterior probabilities of Bayesian analysis; Middle: bootstrap values of maximumparsimony and Right: bootstrap values of maximumlikelihood). These analyses based on the nuclear ITS sequences suggest that all sampled diploid and tetraploid individuals of A. przewalskianum can be grouped into five lineages, AE.
Fig. S3 CHS phylogenetic relationships of sampled Allium przewalskianum in the present study using Bayesian, ML and MP analyses. All analyses were performed according to the methods outlined in the main paper. We used Allium cepa as the outgroup. All newly obtained CHS sequences were deposited in GenBank (accession numbers KC188957–KC189068). We recovered 75 genotypes from 28 populations (Table S7). The ML tree and the MP tree were congruent with the 50% major consensus tree obtained from Bayesian analysis (shown in this Fig., under the GTR+I+G model) and had a similar topology with respect to the main branches. The support values are given above the branches (Left: posterior probabilities of Bayesian analysis; Middle: bootstrap values of maximumparsimony and Right: bootstrap values of maximumlikelihood). These analyses based on the nuclear CHS sequences also suggest that all sampled diploid and tetraploid individuals of A. przewalskianum can be grouped into five lineages, IV.
References:
Cui, X. K., Ao, C. Q., Zhang, Q., Chen, L. T., & Liu, J. Q. (2008). Diploid and tetraploid distribution of Allium przewalskianum Regel. (Liliaceae) in the QinghaiTibetan Plateau and adjacent regions. Caryologia, 61, 192–200.
Evanno, G., Regnaut, S., & Goudet, J. (2005). Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology, 14, 2611–2620.
Pritchard, J. K., Stephens, M., & Donnelly, P. (2000). Inference of population structure using multilocus genotype data. Genetics, 155, 945–959. 