Annals of Human Biology, MayJune 2012; 39(3): 183189

Copyright q Informa UK, Ltd.

ISSN 0301-4460 print/ISSN 1464-5033 online
DOI: 10.3109/03014460.2012.669794

RESEARCH PAPER

Heritability and genetic correlations of obesity-related phenotypes

among Roma people
Alaitz Poveda, Ma Eugenia Ibanez & Esther Rebato
Department of Genetics, Physical Anthropology and Animal Physiology, Faculty of Science and Technology, University of the
Basque Country (UPV/EHU), Bilbao 48080, Spain

infectious disease as the most significant cause of poor

health (Kopelman 2000). The prevalence of obesity has more
than doubled in Western and westernizing countries over
the past decades (James 2004) and according to the World
Health Organization (WHO) it is estimated that, by 2015,
, 2.3 billion adults will be overweight and more than 700
million will be obese. The causes of obesity have both genetic
and environmental components, as shown in several
family (Dasgupta et al. 1997; Raychaudhuri et al. 2003),
twin (Reis et al. 2007; Silventoinen et al. 2008) and adoption
studies (Stunkard et al. 1999). However, the obesity
phenotype is a complex and composite phenotype which
includes various adiposity compartments. Although these
compartments exhibit different biology and hence different
genetic and environmental determination, some common
determinants of the overall degree of fatness may exist
(Hasselbalch et al. 2008). Very few studies have focused on
the existence of pleiotropic effects (i.e. genes simultaneously
influencing several phenotypes) among obesity-related
traits. Livshits et al. (1998), in an attempt to discover the
genetic determinants responsible for the variability and
covariability of the adiposity traits, concluded that
pleiotropic gene effects play an important role in their
variation. In addition to this, Choh et al. (2001) and Mathias
et al. (2009) found pleiotropic effects between BMI and
skinfold thicknesses in Samoan and Danish populations.
Thus, further research is needed in this field to lay down the
basis of the genetic and environmental determinants
responsible for the overall fatness variability.
A promising approach for assessing the contribution
of genetic and environmental factors affecting complex
diseases is the study of population isolates, as these
populations exhibit a relatively uniform genetic background
and low environmental variation (Heutink and Oostra 2002;
Portas et al. 2010). European Gypsies, commonly referred
to as Roma, constitute a genetically isolated population

Background: The Roma people are particularly vulnerable to

developing obesity and related diseases, due to their social
and ethnic backgrounds. However, little is known about the
genetic and/or environmental factors affecting the variability
of obesity-related traits among the Roma population.
Aim: The aim of the present study was to estimate heritabilities
and common genetic and environmental influences of obesityrelated phenotypes in a sample of Roma people living in the
Greater Bilbao region (Basque Country; Spain).
Subjects and methods: Three hundred and seventy-two
individuals from 50 large, extended and highly
consanguineous pedigrees were phenotyped for
anthropometric traits related to obesity. Heritability estimates
were assessed for all quantitative traits and bivariate analyses
were conducted to assess the phenotypic, genetic and
environmental correlations among these traits.
Results: Significant heritable components ( p , 0.01) ranging
from 0.250.68 exist for the studied phenotypes. Heritability
for WHR (h 2 0.60) considerably surpasses the usual
heritability estimates on family-based studies (,0.30).
Measures of overall fatness (BMI, CF and SF) show stronger
correlations with each other than body fat distribution traits
(WHR, CI and TER).
Conclusions: The study concluded that the Greater Bilbao Roma
population is genetically predisposed to abdominal fat
distribution. Variation in body mass is highly associated with
variation in adiposity. However, overall fatness and adiposity
distribution does not seem to share major common genetic
factors, although common environmental factors operate
between them.
Keywords: Adiposity, pleiotropy, ethnic groups, family studies

INTRODUCTION
In recent decades, obesity has reached epidemic proportions
in many developed countries, replacing under-nutrition and

Correspondence: Alaitz Poveda, PhD student, Department of Genetics, Physical Anthropology and Animal Physiology, Faculty of Science and
Technology, University of the Basque Country, Bilbao 48080, Spain, Tel: 34 94 6015405, Fax: 34 94 601 3145. E-mail: alaitz.poveda@ehu.es
(Received 26 July 2011; revised 9 February 2012; accepted 20 February 2012 )

183

184

A. POVEDA ET AL .

which, in contrast to other isolated populations, has

received little attention from the scientific community
(Kalaydjieva et al. 2005). The Roma people originated in
India, possibly in the modern Indian state of Rajasthan or
in the Punjab Region, but it was not until the 14th century
that the Roma population arrived in Europe (Kalaydjieva
et al. 2005). Nowadays, they represent one of the largest
ethnic minorities in Europe, but also one of the most
excluded, experiencing a wide range of social and health
inequalities. As a consequence, Roma people have
considerably poorer health status than the majority
population and are considered to be a high risk group for
obesity and cardiovascular diseases (Vozarova de Courten
et al. 2003; FSG 2009). The incorrect lifestyle of the Romany
minority, who have recently experienced a rapid transition
from a traditional lifestyle to a more sedentary lifestyle with
higher caloric intake (i.e. Westernized lifestyle), may have
played an important role in the increased prevalence of
obesity and related diseases (Simko and Ginter 2009).
These particular features of the Roma population offer a
unique opportunity to understand the role of genetic
predisposition and environmental influences on the risk
of developing obesity. Therefore, in the present study, we
estimate heritabilities, as well as genetic and environmental
correlations for a number of obesity-related traits in a
Roma people sample and compare these results with those
from other populations. To the best of our knowledge, this
represents the first attempt to use quantitative genetic
methods to estimate the heritability of obesity phenotypes
and their covariation in Roma nuclear families.

Table I. Number of relationships between individuals.

Relative pair
Parent offspring
Siblings
Grandparentgrandchild
Avuncular
Half-siblings
Great grandparent grandchild
Grand avuncular
First cousins
First cousins, once removed
First cousins, twice removed
Second cousins
Second cousins, once removed
Double first cousins
Double first cousins, once removed
Double second cousins
Other
Total

Whole
pedigrees (n)

Phenotyped
individuals (n)

784
283
524
424
10
208
128
231
174
11
54
12
3
23
11
61
2941

239
169
53
153
6
4
16
172
117
9
44
11
3
22
9
17
1044

The data collection was carried out in marketplaces,

health centres and public schools with high Roma people
attendance and in family houses of Roma people, during
2009 2011. The idiosyncrasy of the Roma population
makes the access to this population very difficult; therefore,
a tight collaboration with a Roma people association (Kale
Dor Kayiko; KDK) was established during the period of
subject recruitment (2009 2011) to increase the participation of the Roma people. Written informed consent was
obtained from all study participants and permission to carry
out the study in the public centres was requested from the
head of each centre. The project was approved by the ethics
committee of the University of the Basque Country.

METHODS
Sample
The total number of individuals in this study is 655, of whom
372 individuals were phenotyped. Of the 372 measured
individuals, 354 belong to 50 extended pedigrees, including
some very complex four-generation pedigrees. Only
individuals that self-identified as Roma people and who
were living in the Greater Bilbao region (Basque Country,
Spain) were included in the study. In order to avoid possible
admixture of different genetic backgrounds and to maintain
the homogeneity of the sample, only Roma people with
Spanish origin were included in the sample; Rumanian
origin Roma people were excluded in the sample recruitment.
The sample was randomly ascertained, the studied families
were not selected for any specific feature or trait.
In order to avoid the confounding of disease status
and/or medical treatment on the quantitative genetic
estimations, subjects having diagnosed cardiovascular
disease, hyperthyroidism, hypothyroidism or having
experienced gastrointestinal surgery were excluded. The
number of each type of relative pair is listed in Table I.
The overall sample includes 784 parent offspring pairs,
283 sibling pairs, 424 avuncular (aunt/uncle-nice/nephew)
pairs and 231 first cousins pairs in addition to other more
distant relatives.

Phenotypic data
Using standard anthropometric techniques (Lohman et al.
1988), measurements of height, weight, five circumferences
(upper arm relaxed and contracted, waist, hip and medial
calf) and six skinfold thicknesses (biceps, triceps, subscapular, suprailiac, abdominal and medial calf) were
taken from each participant of the study. Height was
measured with a Siber-Hegner anthropometer (GPM,
Zurich, Switzerland) accurate to 1 mm and a digital balance
(accuracy of 0.1 kg) was used to measure body weight. Skinfolds were measured using a Lange caliper (Cambridge
Scientific Industries, Cambridge, MA) and circumferences
by using a Harpenden anthropometric tape (Holtain Ltd.,
Crymych, Pembrokeshire, UK).
From these anthropometric measurements four adiposity-related derived measurements were calculated;
body mass index [BMI weight (kg)/height (m2)],
which is widely considered an overall body mass indicator
and three body fat distribution indices: the waist-tohip ratio (WHR waist circumference/hip circumference),
the trunk-to-extremity skinfold ratio [TER (suprailiac
subscapular abdominal)/(medial calf biceps triceps)]
and the conicity index (CI). This index was calculated
according to Valdez et al. (1993) by using the following
Annals of Human Biology

QUANTITATIVE GENETICS OF OBESITY IN ROMA PEOPLE

p
formula: CI waist circumference (m)/(0.109) [weight
(kg)/height (m)].
While WHR and CI are indices traditionally employed
to assess abdominal obesity, TER opposes trunk and
extremities fat without regard specifically to abdominal fat.
CI quantifies the deviation from the circumference of an
imaginary cylindrical shape. The CI ranges from 1.00
(perfect cylinder) to 1.73 (perfect double cone) (Yasmin and
Mascie-Taylor 2000); therefore, the more central a person is
in fat distribution, the higher the conicity index (Mueller
et al. 1996). Unlike WHR, CI takes into account overall
obesity as modelled from the height and weight of the
individual, independent of hip circumference, which could
be an advantage when comparing men and women who vary
in bone size (Wardle et al. 1996) and in the accumulation of
intra-abdominal fat mass (Yasmin and Mascie-Taylor 2000).
Statistical analysis
Basic descriptive statistics (mean and standard deviation)
were computed for all traits separated by sex and age
group and Students t-test was used to evaluate differences in
mean between males and females in each age group.
Two factor analyses were carried out using the principal
components extraction method for the circumference and
skinfold categories in order to reduce the problem of
multiple comparisons and redundancy of information
(Livshits et al. 2002). The eigenvalue of one criterion was
implemented to retain the factors and scores for each
individual on the extracted two factors (CF and SF) were
used in further analysis.
Next, a stepwise multiple regression analysis was
implemented to remove the effect of age (age, age2, age3)
and sex on all anthropometric traits. In order to improve the
adjustment, the sample was divided by sex and by two age
classes: minors and adults (taking 18 years as the limit for
the two classes). Multiple determination coefficients, R 2, of
these models were used as estimates of the amount of total
trait variance attributable to a significant covariates effect
(Livshits et al. 2005). Phenotypes were then generated for
each individual by retaining the residual regression score
and then standardizing to a mean of 0 and a variance of 1
within each group. After adjustment for age and sex, a
kurtosis test was applied to verify the normality of the
sample distribution and, as none of the traits showed a
kurtosis above 1.96, no transformation was applied to
the data (Blangero et al. 2001). All statistical computations
were carried out using SPSS, versions 18.0 for Windows
(SPSS Inc., Chicago, IL).
Quantitative genetic analysis
In order to quantify the relative additive genetic
contribution to the variation of the analysed traits, a
variance component analysis was performed using the
SOLAR programme (Sequential Oligogenic Linkage Analysis Routines), available online (http:/www.sfbr.org/sfbr/
public/software/solar/solar.html) (Almasy and Blangero
1998). This analysis partitions the observed phenotypic
variance (VP) into additive genetic (VG) and environmental
q Informa UK, Ltd.

185

components (VE), by maximum-likelihood methods

(VP VG VE). The portion of the total phenotypic
variance accounted for by the additive genetic variance
is denoted by narrow sense heritability (h 2); h 2 VG/VP.
The environmental component includes environmental
factors, the non-additive genetic component and measurement errors. Significance of heritability estimates was
tested formally by comparing the likelihoods for the
restricted model, in which additive genetic variance was
constrained to zero, to the likelihood for the general model,
in which additive genetic variance was estimated.
Finally, a bivariate genetic analysis was carried out also in
the SOLAR programme as an extension of the univariate
procedure to determine the extent to which shared genetic
and environmental effects influence the covariation
among pairs of phenotypes. This analysis decomposes
the phenotypic correlations (rP) between pairs of traits into
the underlying genetic (rG) and environmental correlation
(rE), correcting for the use of related individuals, by the
following equation:
qq
p p
rP rG h 2 1 h 2 2 rE 1 2 h 21 1 2 h 22
where h 21 and h 22 are the heritabilities of trait 1 and trait 2,
respectively (Falconer and Mackay 1996). A more detailed
explanation of this methodology can be found in Almasy
and Blangero (1998). Significance of rP, rG and rE between
any pair of traits was tested by comparison of the
log-likelihood values of the full models to a restricted
model, where the parameter of interest is fixed to zero.
Complete pleiotropy was assessed by comparing the general
model, against a restricted model in which rG was
constrained to 1.0 or 2 1.0.
RESULTS
Descriptive statistics
Basic descriptive statistics for all the anthropometric traits
and derived variables separated by sex and age group are
shown in Table II. There were significant differences between
sexes for many traits. Men were significantly taller and
heavier than women, whereas women and girls showed
higher adiposity than men and boys. This is reflected in
the higher mean values observed for women and girls in
skinfolds thicknesses (except for abdominal skinfold).
Significant differences were also found between men and
women for adiposity distribution traits (WHR, TER, CI),
their values being higher for men than for women.
According to the WHO parameters, the present Roma
population is classified as obese, as indicated by the mean
BMIs for both sexes.
Factor decomposition analyses
The results of factor decomposition analyses are presented
in Table III. The KMO (Kaiser-Mayer-Olkin value) was
higher than 0.6 and the Bartletts test was significant
( p , 0.001) in both factors, indicating a good adequacy of

186

A. POVEDA ET AL .
Table II. Phenotypic characteristics of participants and results of Students t-test.
Adults

Traits

Men (n 81)

Age
Height (cm)
Weight (kg)
Arm rel. c. (cm)
Arm cont. c. (cm)
Waist c. (cm)
Hip c. (cm)
Medial calf c. (cm)
Biceps s. (mm)
Triceps s. (mm)
Subscapular s. (mm)
Suprailiac s. (mm)
Abdominal s. (mm)
Medial calf s. (mm)
BMI (kg/m2)
WHR
TER
CI

34.90
171.19
89.52
33.83
36.04
99.54
106.87
38.20
15.65
17.47
24.20
22.30
40.38
21.67
30.59
0.93
1.74
1.26

(11.35)
(6.65)
(18.92)
(4.6)
(4.70)
(15.62)
(9.97)
(3.59)
(8.68)
(8.27)
(8.25)
(9.24)
(14.82)
(11.35)
(6.68)
(0.08)
(0.50)
(0.08)

Women (n 134)

Minors
t-test

Boys (n 66)

Girls (n 91)

t-test

34.27 (11.59)
157.35 (5.78)
75.62 (17.11)
33.48 (5.61)
34.66 (5.64)
87.81 (13.79)
108.87 (12.58)
38.08 (4.32)
25.49 (11.78)
31.82 (11.33)
25.18 (10.10)
27.72 (13.48)
39.21 (14.10)
37.20 (12.25)
30.55 (6.89)
0.81 (0.07)
0.98 (0.22)
1.17 (0.08)

n.s.
10.05 (3.67)
11.59 (3.83)
n.s.
139.76 (20.38)
144.35 (18.41)
n.s.
***
42.60 (18.87)
48.69 (19.85)
n.s.
***
n.s.
24.05 (5.23)
25.87 (5.47)
*
n.s.
25.12 (5.55)
26.55 (5.40)
n.s.
69.50 (10.96)
69.09 (11.26)
n.s.
***
n.s.
80.41 (14.62)
87.52 (16.26)
**
n.s.
30.11 (5.10)
32.58 (5.64)
**
12.50 (7.94)
15.02 (7.24)
***
*
15.85 (8.30)
20.30 (7.95)
***
**
n.s.
12.23 (6.36)
15.55 (7.89)
**
14.30 (10.03)
19.33 (10.41)
**
**
n.s.
21.12 (15.83)
26.50 (15.03)
*
18.80 (10.50)
25.98 (11.14)
***
***
n.s.
20.65 (4.37)
22.32 (5.45)
*
0.87 (0.05)
0.80 (0.07)
***
***
0.98 (0.30)
0.97 (0.23)
n.s.
***
1.18 (0.05)
1.13 (0.06)
***
***
Data are given in mean (standard deviation). N, number of individuals; b, breadth; c, circumference; s, skinfold; n.s., not significant. *p , 0.05,
**p , 0.01, *** p , 0.001.

the sample to the analyses (Suhr 2006). A single factor was

retained for each set of measures. The circumferences (CF)
and skinfold (SF) factors accounted for 91.97% and 80.20%
of the total variance in each group of traits, reflecting the
overall variation of body circumferences and skinfolds,
respectively. Thus, CF could be considered as a measure of
overall body mass and SF as an adiposity indicator. Both
synthetic traits were used as summary variables in the
subsequent quantitative genetic analyses.
Univariate analyses
Heritability estimates (h 2) with associated standard errors
for the obesity-related phenotypes are given in Table IV. The
proportion of variance attributable to age effect (age, age2
and age3) varied considerably depending on trait and age or
sex group, ranging from an R 2 of 0.08 for weight in women
to 0.75 for weight in boys. Covariates explained a higher
proportion of the variance in the minor age group (0.21
0.75) than in adults (0.08 0.51), reflecting the differences in
growth stages among the individuals in the minor age
group. After accounting for the significant covariates effects,
heritability estimates were, in general, highly significant
( p , 0.01), with moderate-to-high values ranging from
0.25 0.60. WHR presented the highest heritability (0.60)
followed by BMI, weight, CF and CI, which showed
heritability estimates of 0.50 or higher and adiposity
indicators (0.46 and 0.25 for SF and TER, respectively).
Bivariate analyses
The phenotypic, genetic and environmental correlations for
obesity-related traits are listed in Table V. Eleven out of 21
phenotypes shared significant genetic effects and all of them
were significantly different from 1.0 or 2 1.0, indicating
incomplete pleiotropy (i.e. shared and unique sets of genes
influenced these traits). CF was the trait that showed the
greatest genetic and environmental correlations with the rest
of measurements. The measures of overall fatness (BMI, CF

and SF) were all highly correlated to each other (rG $ 0.87),
whereas fat distribution traits (WHR, TER and CI) differed
in the mode of inter-relation, with significant and nonsignificant and low-to-high correlations. BMI was the only
overall fatness trait that showed a significant genetic
correlation with all fat distribution indicators. The number
and strength of environmental correlations between overall
body fat traits and fat distribution traits were higher than
the genetic correlations. Among body distribution traits,
whereas WHR and CI showed evidence of shared
environment with all the traits, TER did not show any
environmental correlation with any phenotype.

DISCUSSION
Isolated populations constitute a usually employed resource
for the analysis of the genetic architecture of quantitative
traits (Portas et al. 2010). In the present study, high levels
of statistical significance were obtained with even low
heritabilities and a number of significant correlations
were detected in a genetically isolated Roma population.
Consequently, the complex and unique structure of these
Roma people pedigrees seems to provide valid and robust
Table III. Factor analysis of circumference and skinfold measurements.
Circumference

Factor
loading

Arm relaxed
Arm contracted
Waist
Hip
Medial calf

0.97
0.98
0.92
0.97
0.95

Eigenvalue
% of total variance
KMO (Kaiser-Meyer-Olkin)
Bartletts test

4.60
91.97
0.86
0.00

Skinfold
Biceps
Triceps
Subscapular
Suprailiac
Abdominal
Medial calf

Factor
loading
0.93
0.92
0.86
0.90
0.88
0.89
4.81
80.20
0.89
0.00

Annals of Human Biology

QUANTITATIVE GENETICS OF OBESITY IN ROMA PEOPLE

Table IV. Heritability estimates (h 2), associated standards errors (S.E.)
and proportion of variance attributed to covariate effects (R 2) for
obesity-related phenotypes.
Proportion of variance attributed
to covariate effects (R 2)
Phenotypes

Men

Women

Boys

Girls

h2

S.E.

p-value

Weight
CF
SF
BMI
WHR
TER
CI

0.21
0.23
0.15
0.25
0.50

0.08
0.10
0.14
0.10
0.22

0.75
0.59

0.60
0.52
0.21
0.24
0.55

0.51

0.30

0.50
0.52
0.46
0.54
0.60
0.25
0.52

0.10
0.09
0.11
0.09
0.10
0.11
0.11

, 0.001
, 0.001
, 0.001
, 0.001
, 0.001
0.006
, 0.001

0.31
0.32
0.41
0.26

0.36

Only significant values are reported.

information about the genetic and environmental contribution to variation in obesity-related traits. Our data are
in agreement with previous studies, where heritability of
obesity-related phenotypes were estimated in Indians
(Mathias et al. 2009; Zabaneh et al. 2009), Western
Eurasians (Hasselbalch et al. 2008; Poveda et al. 2010;
Jelenkovic et al. 2011), Mexicans (Bastarrachea et al. 2007)
and Arabs (Bayoumi et al. 2007). The results of this study are
also comparable to those reported in other overweight
populations such as Samoans (Choh et al. 2001) and
Mexican Americans (Comuzzie et al. 1994). Heritability of
body weight in the present sample is very close to the values
reported in American Indian (Choh et al. 2001), Spanish
(Sanchez-Andres and Mesa 1994; Jelenkovic et al. 2011),
Dutch (Zillikens et al. 2008) and Nigerian (Luke et al. 2001)
populations. Concerning BMI, the most extensively used
index for obesity assessment, cross-sectional twin and family
studies have shown a moderate-to-substantial genetic
component in its variation (Coady et al. 2002). The results
of the present study suggest that BMI is significantly
influenced by additive genetic factors, with a heritability
estimate of 50%, which is consistent with the results
reported by other studies using family data (Luke et al. 2001;
Zillikens et al. 2008; Mathias et al. 2009).
The most interesting fact is the high influence of additive
genetic factors on WHR found in this Roma sample
(h 2 60%) which considerably surpass the usual heritability estimates in family-based studies (h 2 , 30%) (e.g.
Mathias et al. 2009; Zabaneh et al. 2009). The high
heritability found for WHR is indicative of a highly heritable
fat distribution pattern in relation to abdominal obesity
among the Greater Bilbao Roma population. In addition,
the high mean values observed for WHR and waist
circumference (see Table II) in this population suggests
that the studied Roma people are genetically pre-disposed to
abdominal fat accumulation. This fact could reflect a
different environmental effect on genes that control
abdominal fat accumulation in the Roma population or it
may also indicate differences in allelic frequencies of these
genes. The rapid transition to a westernized lifestyle and
their particular social structure of small and endogamous
groups may have played an important role in the greater predisposition to abdominal obesity. Although BMI is the most
q Informa UK, Ltd.

187

commonly used indicator of obesity, abdominal fat

accumulation is considered to be the principal component
explaining obesity-related health risks. In fact, WHR has
been frequently viewed as a better predictor of cardiovascular disease risk than BMI in adults (Lee et al. 2010).
Abdominal fat increases the risk of insulin resistance
and cardiovascular disease and is positively correlated
with mortality, even within normal ranges of BMI
(Hasselbalch et al. 2008; Mathias et al. 2009). Thus, the
present sample of Roma people could help in the
identification of specific alleles or environmental effects
acting on the accumulation of abdominal fat which
have important consequences for the development of
cardiovascular and other obesity-associated diseases.
Phenotypic correlations among the great majority of
obesity-related traits are shown in this sample. Pleiotropy
was also found for an elevated number of pairs. The
phenotypic, genetic and environmental correlations estimated from the present study are mostly in line with
previous studies (Choh et al. 2001; Schousboe et al. 2004;
Bastarrachea et al. 2007; Hasselbalch et al. 2008; Mathias
et al. 2009). The high phenotypic, genetic and environmental correlation for BMI-SF reported in the present study,
which is in agreement with the studies of Hasselbalch
et al. (2008) and Schousboe et al. (2004), indicate that
variation in body mass is highly associated with variation in
adiposity. Livshits et al. (1998), in a study conducted on
three ethnically different populations, found that skinfold
indices measure a different dimension of fat distribution
than circumference indices, being to a great extent
genetically independent characteristics. However, the present study shows a high genetic correlation (0.87) between
CF and SF, pointing to the high influence of pleiotropic
effects on the determination of both circumference and
skinfold measurements.
Table V. Phenotypic (rP), genetic (rG) and environmental (rE)
correlations between obesity-related phenotypes.
Weight
CF
SF
BMI
WHR
TER
rP
0.95a
rG
0.96a
rE
0.93a
SF
rP
0.84a
0.86a
rG
0.80a
0.87a
a
rE
0.88
0.85a
BMI
rP
0.93a
0.96a
0.86 a
a
a
rG
0.92
0.96
0.87a
rE
0.93a
0.97a
0.86a
a
a
WHR rP
0.29
0.28
0.32a
0.37a
rG
0.22
0.17
0.15
0.32a
b
b
a
rE
0.38
0.42
0.50
0.44b
TER
rP
0.11
0.10
0.10
0.16b
0.18b
c
rG
0.42
0.31
0.35
0.48
0.48c
rE 20.07 20.02 20.03
20.02 20.01
rP
0.37a
0.37a
0.41a
0.42a
0.86a 0.21a
CI
rG
0.31
0.25
0.24
0.36c
0.96a 0.43
b
a
a
a
rE
0.44
0.51
0.58
0.49
0.75a 0.09
a
p-values , 0.05 are marked in italics; Estimated significant at
CF

p , 0.001; b Estimated significant at p , 0.01; c Estimated significant at

p , 0.05.

188

A. POVEDA ET AL .

In general, measurements of overall fatness (BMI, SF6,

CF and SF) show stronger correlations with each other than
body fat distribution traits (WHR, CI and TER), indicating
a higher influence of shared genetic and environmental
factors on overall fatness than on adiposity distribution. The
stronger environmental than genetic correlations between
overall body fat traits and fat distribution traits show that,
although common environmental factors, such as energy
intake, operate among obesity-related traits, overall fatness
and adiposity distribution do not seem to share major
common genetic factors. Although BMI is weakly correlated
with WHR, it shows high genetic and environmental
correlations (, 0.90) with waist and hip circumferences
independently (data not shown), as observed in previous
studies (Bastarrachea et al. 2007; Benyamin et al. 2007;
Hasselbalch et al. 2008; Mathias et al. 2009). Thus, this fact
suggests that while overall body fatness is strongly related to
fat accumulation in body regions, overall body fatness is
only weakly correlated with fat distribution.
CONCLUSION
In conclusion, our findings have provided evidence for a
strong genetic contribution to the variation of obesity-related
phenotypes among the Greater Bilbao Roma population
which is genetically pre-disposed to abdominal obesity.
Pleiotropic effects have a great influence among various
obesity-related traits, but none of the studied obesity-related
traits pairs show complete pleiotropy, indicating a significant
residual genetic influence specific for traits. In the present
Roma people sample, overall fatness and adiposity
distribution do not seem to share major common genetic
factors although common environmental factors operate
among them. Understanding the complex relationships
between the underlying genetic and environmental factors in
the variation of obesity phenotypes will help provide insights
into the complex aetiology of obesity. Further study of these
families may provide useful information for continuing
genetic research in this interesting anthropological population, including the identification of potentially unique
genetic loci in Roma people.
ACKNOWLEDGEMENTS
We would like to thank the outpatient departments of
health centres and the education centres who allowed us to
carry out the study. We also would like to express our
gratitude to Kale Dor Kayiko for their collaboration. Special
thanks also go to the families enrolled in the study for their
participation.
Declaration of interest: This study was supported by grants
of the Bilbao Bizkaia Kutxa (BBK; 87014/97012/07007), of
the Spanish Ministry of Science and Innovation (MICINN;
GCL2010-15511), of the Industry Department of the Basque
Government (SAIOTEK; SA2010/00035) and by two predoctoral grants one of the Ministry of Education of Spain
(for A.P) and the other from the University of the Basque

Country (for M.E.I.). The authors report no conflicts of

interest. The authors alone are responsible for the content
and writing of the article.

REFERENCES
Almasy L, Blangero J. 1998. Multipoint quantitative-trait linkage
analysis in general pedigrees. Am J Hum Genet 62:1198 1211.
Bastarrachea RA, Kent JW, Jr, Rozada G, Cole SA, Lopez-Alvarenga JC,
Aradillas C, Brito-Zurita O, Cerda-Flores RM, Ibarra-Costilla E,
Gallegos E, Laviada-Molina H, Hernandez-Escalante V, Rosas J,
Machado A, Vadillo F, Ramos M, Lazalde B, Santa-Olalla J,
MacCluer JW, Comuzzie AG. 2007. Heritability and genetic
correlations of metabolic disease-related phenotypes in Mexico:
preliminary report from the GEMM Family Study. Hum Biol Int
Record Res 79:121 129.
Bayoumi RA, Al-Yahyaee SA, Albarwani SA, Rizvi SG, Al-Hadabi S, AlUbaidi FF, Al-Hinai AT, Al-Kindi MN, Adnan HT, Al-Barwany HS,
Comuzzie AG, Cai G, Lopez-Alvarenga JC, Hassan MO. 2007.
Heritability of determinants of the metabolic syndrome among
healthy Arabs of the Oman family study. Obesity 15:551 556.
Benyamin B, Sorensen TI, Schousboe K, Fenger M, Visscher PM, Kyvik
KO. 2007. Are there common genetic and environmental factors
behind the endophenotypes associated with the metabolic
syndrome? Diabetologia. 50:1880 1888.
Blangero J, Williams JT, Almasy L. 2001. Variance component methods
for detecting complex trait loci. Adv Genet 42:151 181.
Choh AC, Gage TB, McGarvey ST, Comuzzie AG. 2001. Genetic and
environmental correlations between various anthropometric and
blood pressure traits among adult Samoans. Am J Phys Anthropol
115:304 311.
Coady SA, Jaquish CE, Fabsitz RR, Larson MG, Cupples LA, Myers RH.
2002. Genetic variability of adult body mass index: a longitudinal
assessment in framingham families. Obes Res 10:675 681.
Comuzzie AG, Blangero J, Mahaney MC, Mitchell BD, Stern MP,
MacCluer JW. 1994. Genetic and environmental correlations among
skinfold measures. Int J Obes Relat Metab Disord 18:413 418.
Dasgupta I, Dasgupta P, Daschaudhuri AB. 1997. Familial resemblance
in height and weight in a endogamous mahisya caste population of
rural west Bengal. Am J Hum Biol 9:7 9.
Falconer DS, Mackay TFC. 1996. Introduction to Quantitative Genetics.
Fourth edition. Addison Wesley Longman, Harlow, Essex, UK.
FSG. 2009. Health and the Roma Community, analysis of the situation
in Europe. Bulgaria, Czech Republic, Greece, Portugal, Romania,
Slovakia, Spain. Published by: EFXINI POLI - Local Authorities for
Social, Cultural, Tourist, Environmental and Agricultural Development. Greece. FSG - Fundacion Secretariado Gitano. Spain.
Office of the council for Roma community affairs. Czech Republic.
PDCS-Partners for Democratic Change Slovakia. REAPN - Rede
Europeia Anti-Pobreza/Portugal. ROMANI CRISS - Roma Center
for Social Interventions and Studies. Romania. THRPF - The Health
of Romany People Foundation. Bulgaria. Coordinated by:
Fundacion Secretariado Gitano Health Area. Edited by: Fundacion
Secretariado Gitano.
Hasselbalch AL, Benyamin B, Visscher PM, Heitmann BL, Kyvik KO,
Sorensen TI. 2008. Common genetic components of obesity traits
and serum leptin. Obesity 16:2723 2729.
Heutink P, Oostra BA. 2002. Gene finding in genetically isolated
populations. Hum Mol Genet 11:2507 2515.
James PT. 2004. Obesity: the worldwide epidemic. Clin Dermatol 22:
276 280.
Jelenkovic A, Poveda A, Rebato E. 2011. Quantitative genetics of
human morphology and obesity-related phenotypes in nuclear
families from the Greater Bilbao (Spain): comparison with other
populations. Ann Hum Biol 38:471 478.
Kalaydjieva L, Morar B, Chaix R, Tang H. 2005. A newly discovered
founder population: the Roma/Gypsies. Bioessays 27:1084 1094.
Annals of Human Biology

QUANTITATIVE GENETICS OF OBESITY IN ROMA PEOPLE

Kopelman PG. 2000. Obesity as a medical problem. Nature 404:
635 643.
Lee J, Chen L, Snieder H, Chen da F, Lee LM, Liu GF, Wu T, Tang X,
Zhan SY, Cao WH, Lv J, Gao WJ, Hu YH. 2010. Heritability of
obesity-related phenotypes and association with adiponectin gene
polymorphisms in the Chinese national twin registry. Ann Hum
Genet 74:146 154.
Livshits G, Pantsulaia I, Gerber LM. 2005. Association of leptin levels
with obesity and blood pressure: possible common genetic
variation. Int J Obes 29:85 92.
Livshits G, Roset A, Yakovenko K, Trofimov S, Kobyliansky E. 2002.
Genetics of human body size and shape: body proportions and
indices. Ann Hum Biol 29:271 289.
Livshits G, Yakovenko K, Ginsburg E, Kobyliansky E. 1998. Genetics of
human body size and shape: pleiotropic and independent genetic
determinants of adiposity. Ann Hum Biol 25:221 236.
Lohman TG, Roche AF, Martorell R. 1988. Anthropometric
standardization reference manual. Champaign, IL: Human Kinetics
Books.
Luke A, Guo X, Adeyemo AA, Wilks R, Forrester T, Lowe W, Jr,
Comuzzie AG, Martin LJ, Zhu X, Rotimi CN, Cooper RS. 2001.
Heritability of obesity-related traits among Nigerians, Jamaicans
and US black people. Int J Obes Relat Metab Disord 25:1034 1041.
Mathias RA, Deepa M, Deepa R, Wilson AF, Mohan V. 2009.
Heritability of quantitative traits associated with type 2 diabetes
mellitus in large multiplex families from South India. Metab Clin
Exp 58:1439 1445.
Mueller WH, Meininger JC, Liehr P, Chan WY, Chandler PS. 1996.
Conicity: a new index of body fat distribution - What does it tell us?
Am J Hum Biol 8:489 496.
Portas L, Murgia F, Biino G, Concas MP, Casula L, Milia S, Whalen MB,
Vaccargiu S, Cosso M, Parracciani D, Frongia B, Pirastu M. 2010.
History, geography and population structure influence the
distribution and heritability of blood and anthropometric
quantitative traits in nine Sardinian genetic isolates. Genet Res 92:
199 208.
Poveda A, Jelenkovic A, Susanne C, Rebato E. 2010. Genetic
contribution to variation in body configuration in Belgian nuclear
families: a closer look at body lengths and circumferences.
Collegium Antropol 34:515 523.
Raychaudhuri A, Ghosh R, Vasulu TS, Bharati P. 2003. Heritability
estimates of height and weight in Mahishya caste population. Int J
Hum Genet 3:151 154.
Reis VM, Machado JV, Fortes MS, Fernandes PR, Silva AJ, Dantas PS,
Filho JF. 2007. Evidence for higher heritability of somatotype
compared to body mass index in female twins. J Physiol Anthropol
26:9 14.

q Informa UK, Ltd.

189

Sanchez-Andres A, Mesa MS. 1994. Heritabilities of morphological and

body composition characteristics in a Spanish population.
Anthropologischer Anzeiger; Bericht uber die biologisch-anthropologische Literatur 52:341 349.
Schousboe K, Visscher PM, Erbas B, Kyvik KO, Hopper JL, Henriksen
JE, Heitmann BL, Sorensen TI. 2004. Twin study of genetic and
environmental influences on adult body size, shape, and
composition. Int J Obes Relat Metab Disord 28:39 48.
Silventoinen K, Magnusson PK, Tynelius P, Kaprio J, Rasmussen F.
2008. Heritability of body size and muscle strength in young
adulthood: a study of one million Swedish men. Genet Epidemiol
32:341 349.
Simko V, Ginter E. 2009. Short life expectancy and metabolic syndrome
in Romanies (Gypsies) in Slovakia. Cent Eur J Public Health 18:
1618.
Stunkard AJ, Berkowitz RI, Stallings VA, Cater JR. 1999. Weights of
parents and infants: is there a relationship? Int J Obes Relat Metab
Disord 23:159 162.
Suhr DD. 2006. Exploratory or confirmatory factor analysis?
Proceedings of the Thirty-first Annual SAS Users Group
International Conference. Cary, NC: SAS Institute. Paper 200-31.
17 pages.
Valdez R, Seidell JC, Ahn YI, Weiss KM. 1993. A new index of
abdominal adiposity as an indicator of risk for cardiovascular
disease. A cross-population study. Int J Obes Relat Metab Disord 17:
7782.
Vozarova de Courten B, de Courten M, Hanson RL, Zahorakova A,
Egyenes HP, Tataranni PA, Bennett PH, Vozar J. 2003. Higher
prevalence of type 2 diabetes, metabolic syndrome and cardiovascular diseases in gypsies than in non-gypsies in Slovakia.
Diabetes Res Clin Pract 62:95 103.
Wardle J, Wrightson K, Gibson L. 1996. Body fat distribution in South
Asian women and children. Int J Obes Relat Metab Disord 20:
267 271.
Yasmin, Mascie-Taylor CGN. 2000. Adiposity indices and their
relationship with some risk factors of coronary heart disease
in middle-aged Cambridge men and women. Ann Hum Biol 27:
239 248.
Zabaneh D, Chambers JC, Elliott P, Scott J, Balding DJ, Kooner JS.
2009. Heritability and genetic correlations of insulin resistance
and component phenotypes in Asian Indian families using a
multivariate analysis. Diabetologia 52:2585 2589.
Zillikens MC, Yazdanpanah M, Pardo LM, Rivadeneira F, Aulchenko
YS, Oostra BA, Uitterlinden AG, Pols HA, van Duijn CM. 2008.
Sex-specific genetic effects influence variation in body composition.
Diabetologia 51:2233 2241.

Copyright of Annals of Human Biology is the property of Taylor & Francis Ltd and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.