Mediterranean Y Chromosome Population Structure

656
87
55
70
68
81
90
39
19
39
43
65
on hgs frequencies, using the same 8 hgs tested for PC

n
(2)
(1)
(2)
(2)

0.033 (3)
0.026 (1)

0.026 (1)

0.031 (2)
analysis. The pattern of population similarities paral-

0.023
0.018
0.029
0.029
R1∗ (xR1a1) R1a1
leled that displayed by PC analysis, with populations in
the same cluster being mostly not significantly different
from each other (P > 0.05, data not shown). Population

(17)
(16)
(18)
(17)
(17)
(29)
(2)

0.051 (2)
(3) 0.047 (2)
0.092 (6)
relationships were also investigated by the use of mi-

0.195
0.291
0.257

0.322
0.051
0.25
0.21
crosatellite variation on a total of 34 populations (Table

Table 3 Population Y hg frequencies (in brackets absolute number of chromosomes) investigated in this study. No hg D and N3 chromosomes were found
1). Analysis of the distribution of molecular variation on

(1)
(1)
(2)
(1)
P∗ (xR1)
STRs haplotypes (performed using Arlequin software;

0.014
0.015
0.025
0.011

0.047 (2) 0.07

Schneider et al. (2000)) was also consistent with the four
group clustering suggested by the PC analysis, with the

(4)
(3)
(1)
(2)
(1)
(4)

0.051 (2)

0.046 (3)
lowest within groups variation (1.57%) and largest be-

K∗ (xP,

0.046
0.055
0.014
0.029
0.012
0.044
N3)
tween groups variation (7.70%,
st 0.077, P  0.01).
Clustering of the West and Central-East Mediterranean

(10)
(15)
(4)
(8)
(8)

(6)
(2)
(2)
(5)
(8)
(6)
populations lowered the intergroup variation to 6.95 %

F∗ (xI,

0.046
0.145
0.114
0.147
0.185
0.067
0.051
0.105
0.128
0.186
0.092
and increased the intragroup value to 2.35% (P  0.01).

J,K)
Following a geographical clustering scheme as indicated

(4) 0.259 (21)

(3) 0.011 (1)

(3) 0.015 (1)

in Table 1, genetic variation showed a between groups

I1b2
value of 5.18% and a within groups percentage of 3.06
(P  0.01). Finally, clustering the North African Arab

(11)

(11)
(3)

(1)

0.047 (2)
0.077 (5)
I∗ (xI1b2)
samples with the Arab cluster of Figure 3a displayed a

0.034
0.055
0.157
0.044
0.049
0.122
0.026
between groups value of 6.18% and a within groups
percentage of 2.71 (P  0.01).
(25)
(15)

(11)

(19)

(10)
(10)
(15)
(24)
We also evaluated the degree of differentiation shown
(8)

(8)

(4)
by Mediterranean populations by performing PC anal-
0.287
0.273
0.114
0.162
0.099
0.211
0.103
0.526
0.256
0.349
0.369
J2

ysis on previously published mtDNA haplogroup data

0.308 (12)
(Richards et al. 2002; Thomas et al. 2002; Plaza et al.
0.069 (6)

0.014 (1) 0.071 (5)

0.044 (3)
0.049 (4)
0.078 (7)
0.231 (9)
0.316 (6)

0.093 (4)
0.062 (4)
J∗ (xJ2)

2003). In general agreement with the Y chromosome

results, a Mediterranean, an Arab and a North African
cluster were apparent (Fig. 3b). Contrary to the Y chro-
mosome PC plot (Figure 3a) West and Central Mediter-
ranean samples did not cluster independently, while
E∗ (xE3a, C

(8) 0.011 (1) 0.011 (1)

a certain tendency to separate was shown by Middle

Eastern samples. Mediterraneans and North Africans
E3b)

were closer to each other than to Arabic populations

(Bedouins and Yemenite), a result previously interpreted
as a signature of female mediated gene flow across the
E3a

Mediterranean basin (Plaza et al. 2003). Near Eastern

(25)

(15)
(18)

(20)

(13)
(9)

(8)

(1)
(7)
(7)

Arab populations (Palestinians, Jordanians and Syrians)

did not cluster with the Arabians and were instead closer
0.287
0.164
0.214
0.265
0.012 (1) 0.099
0.089
0.513
0.053
0.179
0.163
E3b

0.031 (2) 0.2

to Mediterranean populations. This result is in contrast

with the Y chromosomal data, where Arabian and non-
Arabian Arabs grouped together (Fig. 3a). Sex mediated
Sample A

differential contacts between the Arabian peninsula and

the Near East could be responsible for the different ob-
NWS

TUN
MAL

MLE

CYP
SWS

SAR

CLE
ESC

DJE
SIT

served pattern. In the Middle East, only Bedouins seem


C University College London 2005 Annals of Human Genetics (2005) 7