Investigation of associations of European, African, Amerindian genomic ancestries and MC4R, FTO, FAIM2, BDNF loci with obesity-related traits in Rio de Janeiro, Brazil

Abstract

A complex web of causation is involved in adiposity, including environmental, social and genetic factors. We aimed to investigate associations between genetic factors such as ancestry and single nucleotide polymorphisms, and obesity-related traits in a sampled Brazilian population. A sample of 501 unrelated adults participating in 2013 at the longitudinal Pró-Saúde Study (EPS) in Rio de Janeiro, Brazil was selected. We analysed 46 AIM-InDels (insertion/deletion) as genetic ancestry markers and four single nucleotide polymorphisms located in the genes MC4R (rs17782313), FTO (rs9939609), FAIM2 (rs7138803) and BDNF (rs4074134), previously described as associated with obesity. The selected obesity-related markers were anthropometric parameters such as body mass index, waist circumference and waist-to-hip ratio, and body composition measurements namely body fat percentage, android fat mass and gynoid fat mass. The sample showed greater European ancestry (57.20%), followed by African (28.80%) and lastly Amerindian (14%). Our results suggest that the rs4074134 (BDNF) CC genotype was directly associated with gynoid fat mass, whereas body fat percentage, android fat mass and the anthropometric parameters seem not to be associated with neither ancestry nor the four polymorphisms in this population sample, most likely due to a stronger role of social, behavioural and environmental determinants.

Key words adiposity; ancestry; association study; Brazilian population; obesity

INTRODUCTION

Obesity, a condition once rare, increased over time in incidence, and was finally identified as a health problem in the middle of the 19th century (Wells 2006). However, since 1980, the Body Mass Index (BMI = weight/height²) has already increased by an average of 0.4 kg/m² per decade worldwide. In 2000, for the first time, the number of overweight adults surpassed those with a deficit (Finucane et al. 2011), and ten years later the worldwide prevalence of obesity has doubled. In 2016, the World Health Organization (WHO) published that 39% of adults were already overweight (BMI ≥ 25 kg/m²) and 13% were obese (BMI ≥ 30 kg/m²) (World Health Organization 2020). Low and middle income countries, largely located in Africa and the Americas, have registered a faster increase in the prevalence of overweight and obesity (Ng et al. 2014).

In Brazil, a national study called Surveillance of Risk and Protection Factors for Chronic Diseases by Telephone Survey identified that 55.40% of individuals aged 18 or over are overweight and 20.30% are obese. Prevalence of overweight increased from 42.60% in 2006 to 55.40% in 2019, being highest among men (57.10%) and adults aged 65 and over (59.80%). Prevalence of obesity was highest among women (21%) and also increased with age, reaching 20.90% among adults aged 65 and over. In Rio de Janeiro, the prevalence of obesity and overweight resemble the national average (Ministério da Saúde 2020).

Obesity is a complex condition influenced by genetic, behavioural, nutritional, social and environmental factors (Loos & Bouchard 2003, Velázquez-Fernández et al. 2017). The most common form of obesity appears not to follow a genetic inheritance model and is caused by an interaction of major effect environmental factors with multiple genetic variants with minor effect (Da Fonseca et al. 2017).

The role of genetics in susceptibility to obesity has been investigated since the early 2000s. The most accessible parameter for the assessment of genetic factors and obesity risk is the BMI, although other parameters (eg. waist circumference, waist-to-hip ratio and body fat percentage) have emerged given its association with obesity and some of its risk factors (World Health Organization 2008, Duclos 2016, Patnaik et al. 2017).

Genome wide association studies (GWAS) is a well-established and effective approach of identifying candidate risk loci in the human genome containing genetic variations associated with complex diseases or phenotypes (Patron et al. 2019). Wu et al. (2018) through a GWAS recently revealed 291 nominally obesity-associated genes in a Northern Han Chinese cohort and showed that these genes have an important role in different pathways, such as homeostasis, olfactory transduction and platelet production. Interestingly, several GWAS have been consistently identified candidate genes that may be causally linked to obesity, such as MC4R, FTO, FAIM2 and BDNF (Hotta et al. 2009, Ng et al. 2010, Dorajoo et al. 2012).

Furthermore, there is also a possible association between ancestry, ethnicity and the clinical syndrome, showing the importance of better understanding the history of populations (Sabin et al. 2011, Waalen 2014, Pigeyre et al. 2016, Yu 2017, Qasim et al. 2018). This aspect may be even more relevant in admixed populations from well-differentiated groups (Grant et al. 2008, 2009, Ningombam et al. 2018). Obesity-related traits have been associated with the European, African or Native ancestries in different ethnic groups. It is interesting to note that some studies demonstrate that the estimates of admixture from an ancestral population are directly related to BMI, while others report an inverse association (Fernández et al. 2013). These discrepancies seem to be affected by ethnicity, environmental factors, gender and the number of ancestral populations considered in the study.

The genetic markers called “Ancestry Informative Markers (AIMs)” have a high differentiation between their allele frequencies in different or geographically distant ancestral populations. These markers are especially useful to infer the ancestral origin of a given individual and to estimate the proportions of different lineal origins of admixed populations (Shriver et al. 1997). In studies of genetic association with obesity risk, ancestry may be a confounder. It is vital to use AIMs as a control variable along with the single nucleotide polymorphisms described as related to obesity in populations that may be under stratification, in order to avoid the observation of erroneous associations.

The aim of this study was to investigate whether FTO, MC4R, FAIM2 and BDNF variants/loci identified in GWAS as well as AIMs contribute to the obesity-related traits in a Brazilian population.

The data that support the findings of this study are available from the corresponding author, upon reasonable request.

MATERIALS AND METHODS

Study design and population

This is a cross-sectional study conducted in working-age unrelated individuals from a cohort study of university civil servants in Rio de Janeiro, the Pró-Saúde Study (EPS). The EPS is a longitudinal prospective study of non-faculty public agents at the State University of Rio de Janeiro, Brazil, which focus on the investigation of health-related social and behavioural determinants (Faerstein et al. 2005). Four waves of data collection have been conducted among 3 253 participants (1999, 2001, 2006, and 2012). During the fourth data collection phase, a sample of 520 participants was selected to perform additional assessments absent in the first three waves, including body composition by dual energy X-ray absorption (DXA) and blood collection for genetic ancestry analysis. This sample was randomly defined considering the proportions of the strata of gender, age (up to 50 years vs. 51 years or more) and level of education (up to complete high school vs. complete university or more) in the cohort population.

All samples were collected under written informed consent and the project was approved by the Research Ethics Committee of the Institute of Social Medicine of the State University of Rio de Janeiro (CAAE 0041.0.259.000-11 and 04452412.0.0000.5260). All procedures performed in this study involving human participants were in accordance with the ethical principles of the 2000 Helsinki Declaration of the World Medical Association (2013).

Samples and DNA extraction

The extraction and purification of leukocyte DNA from total blood samples was performed using the commercial kit (Puregene Blood Kit, Qiagen, Düsseldorf, Germany). DNA quality and integrity were evaluated via electrophoresis on 0.80% agarose gel. DNA samples were quantified by spectrophotometry in NanoDrop (Thermo Fisher Scientific, Carlsbad, CA, USA), aliquoted and stored at -80 °C until use.

Genotyping of 46 AIM-InDels and ancestry estimation

All samples were genotyped for 46 AIM-InDels by PCR and capillary electrophoresis according to the protocol described in Pereira et al. (2012). The 46 AIM-InDels system used was optimized to efficiently measure admixture proportions from four different continental origins (Africa, Europe, America and East Asia) in a single PCR-multiplex (Pereira et al. 2012).

The amplified products were analysed by capillary electrophoresis on an ABI 3500 (Applied Biosystems, Waltham, MA, USA). Genotyping was performed using the software GeneMapper ID-X v1.2 and/or GeneMapper ID v4.1 (Thermo Fisher Scientific).

Ancestry estimates of the individuals from the population sample were obtained using the software STRUCTURE v2.3.4 (Pritchard et al. 2000), with reference population data from 98 African (AFR), 147 European (EUR) and 51 Native (NAM) individuals from the HGDP-CEPH diversity panel as reference (Pereira et al. 2012).

The STRUCTURE analyses were carried out using 100 000 burnin steps, followed by 100 000 iterations Markov Chain Monte Carlo (MCMC). In the analysis, three clusters (K) were assumed, and to verify the consistency of the results, five independent runs were performed. These five runs were combined using the software CLUMPP v.1.1.222 (Jakobsson & Rosenberg 2007) to obtain the best estimate of the cluster membership coefficients.

Genotyping of the four SNPs described in association to obesity

Four single nucleotide polymorphisms (SNPs) described in association to overweight/obesity factors were genotyped in the Brazilian population sample: rs17782313 (MC4R), rs9939609 (FTO), rs7138803 (FAIM2) and rs4074134 (BDNF).

The rs17782313 (MC4R) was the first locus in which mutations directly associated with dominant inherited human obesity were identified, and was the most investigated genetic cause of obesity before the era of GWAS (Huszar et al. 1997, Yeo et al. 1998). The rs17782313 polymorphism has been associated with increased BMI and waist circumference (Loos 2011), insulin resistance (Tschritter et al. 2011), diabetes mellitus type 2 and lipid intake (Khalilitehrani et al. 2015). In addition, the presence of the risk allele (C) has shown a relationship with uncontrolled food intake (Vega et al. 2016).

Described as one of the genetic factors that lead to the onset of obesity and its metabolic implications by the hedonic mechanism of food reward, the rs9939609 (FTO) increased expression may lead to obesity (Church et al. 2010). FTO associations with body fat mass and other obesity-related traits have been identified in different investigations (Dina et al. 2007, Frayling et al. 2007, Scuteri et al. 2007, Shabana & Hasnain 2015).

The rs7138803 (FAIM2) is an apoptosis-inhibiting gene apparently related to obesity. Epigenetic changes are important factors in helping develop obesity and it can help understand this process. Methylation levels at 8 sites in the gene promoter were significantly different between obese and lean individuals. This study provided the first evidence that the methylation levels of the FAIM2 promoter are significantly associated with obesity (Wu et al. 2015).

The least investigated, rs4074134 (BDNF) has been explored by fewer studies, and most of them were in Asian populations. Han et al. (2013) found an association between this polymorphism and the reduction of BMI, WC, glucose, insulin and risk for type 2 diabetes in Chinese Han population. In a Japanese population, however, Hotta et al. (2009) observed that this genetic variation, as well as SNPs in FAIM2 and MC4R genes, was marginally associated with obesity susceptibility. Through GWAS, seven SNPs located within or downstream of BDNF at 11p14, including rs4074134 (BDNF), were found to be associated with BMI and total mass (Thorleifsson et al. 2009).

These variants were determined through real-time polymerase chain reaction TaqMan® SNP genotyping assays (Thermo Fisher Scientific). Amplification was performed in a StepOne® Plus real-time PCR system (Thermo Fisher Scientific) and data analyses were carried out using StepOne software v2.3 (Thermo Fisher Scientific).

Obesity-related markers assessment

The total body mass, total fat mass, and android and gynoid fat masses (AFM and GFM respectively) were measured via DXA with iDXA Lunar equipment (GE Healthcare, Milwaukee, WI, USA) and enCORE 2008 software version 12.20 (iDXA GE Medical Systems enCORE Operator’s Manual and Product Information). The body fat percentage (BF) was derived from the total mass and the total fat mass (total fat mass/total mass x100).

For the full-body examination, participants used light clothing and no accessories that could skew the measurements. They were placed in the supine position and asked to remain immobile until the end of the procedure. The android region was defined by the software as the region from the top of the iliac crest to 20% of distance from chin to iliac crest, and the gynoid region was defined as the upper limit below the pelvis cut line by 1.5 times the height of the android region. The height of the gynoid region was equal to two times the height of the android region. All DXA scans were performed by the same trained professional and followed standard quality control procedures according to the manufacturer. The equipment was calibrated daily according to the manual’s protocol (iDXA GE Medical Systems enCORE Operator’s Manual and Product Information).

For the examination of waist circumference (WC) and hip measurement, in centimeters (cm) the participants were recommended to wear light clothing and without metal accessories, according to the standard protocol (Lohman et al. 1992). The waist-to-hip ratio (WHR) resulted from the division of the waist circumference by the hip measurement.

Height was measured with a wall stadiometer with an accuracy of 0.1 cm, while individuals remained barefoot in the orthostatic position and with their heads oriented in the horizontal position of the Frankfurt plane. The BMI calculation was the result of the total mass and height (kg/m²).

Statistical analysis

The Hardy-Weinberg Equilibrium (HWE) tests, for the ancestry markers and SNPs, were performed using the Arlequin software version 3.5 (Excoffier & Lischer 2010).

Multiple linear regressions were used to examine the associations between each proportion of genetic ancestry and each genotype of the four SNPs with obesity-related measures in the sampled population. Since three female individuals do not have information on BMI and three other female individuals lack information on body fat percentage, and android and gynoid fat masses, the sample size varied in the analyses from 498 for these parameters to 501 for waist circumference and waist-to-hip ratio. Normality of continuous variables was tested by Shapiro Wilk test. All continuous parameters had a normal distribution.

Each ancestry — African, European and Amerindian — was used, through its proportions, as a continuous variable for the multiple linear regression analyses, and as nominal variables for the mean values assessment of the obesity-related traits in the sampled population. When used as nominal variable, AFR, EUR or NAM ancestry of the individuals from the population sample was defined according to the majority ancestry of each individual. To test the associations, six markers related to obesity were used: three anthropometric parameters namely BMI, waist circumference and waist-to-hip ratio, and three body composition measures such as body fat percentage, and android and gynoid fat masses. To avoid spurious associations, the following confounders were included in the analysis between ancestry and the obesity markers as control variables: age, gender and income.

For the analysis between SNP genotypes and the obesity traits, all the previous control variables were used, plus the continuous variable ancestry. Each genotype, TT, CT, CC for rs17782313 (MC4R); TT, AT, AA for rs9939609 (FTO); GG, AG, AA for rs7138803 (FAIM2); and CC, CT, TT for rs4074134 (BDNF), was encoded into a ‘dummy variable’ so that it was possible to use multiple linear regressions (Garavaglia & Sharma 1998, Draper & Smith 1998). The alleles described as having the highest risk associated with obesity and metabolic conditions are, in each SNP: C, A, A and C for rs17782313 (MC4R), rs9939609 (FTO), rs7138803 (FAIM2) and rs4074134 (BDNF), respectively (Howe et al. 2021). The interpretation of the analyses’ results with ‘dummy variables’ coding is in relation to the ‘dummy variables’ of reference, i.e., the heterozygous and homozygous genotypes of the risk alleles were analysed and interpreted in reference to the homozygous genotypes of lowest risk, which are not included in the association tests. Statistical analyses were performed using IBM SPSS software version 20.0 (IBM Corp., Armonk, NY, USA). A significance level of 0.05 was considered.

RESULTS

The descriptive characteristics of the population sample by the obesity-related markers are presented in Table I. The study consisted of 240 men and 261 women. The total average age of the sample was 51.62 years (SD = 7.90), and the total average income was 2246.05 BRL (SD = 1469.20). The average age for men and women was 51.23 years (SD = 7.81) and 51.68 years (SD = 7.98), respectively. The average income for men was 2260.49 BRL (SD = 1538.00) and for women was 2232.78 BRL (SD = 1405.84). Differences in mean values between men and women for the obesity-related traits revealed that variations were not statistically significant for BMI and android fat mass.

Table I
Descriptive characteristics of the population sample containing means and standard deviations of total sample and between men and women for each obesity-related parameter.

	Total / Gender	n	Mean (SD)	p (between men and women)
BMI (kg/m²)	Total Men Women	498 240 258	28.04 (5.05) 27.93 (4.58) 28.15 (5.46)	0.635
WC (cm)	Total Men Women	501 240 261	97.10 (12.47) 98.81 (12.35) 95.53 12.38)	0.003
WHR	Total Men Women	501 240 261	0.94 (0.08) 0.97 (0.06) 0.90 (0.07)	0.001
BF (%)	Total Men Women	498 240 258	36.26 (8.32) 30.42 (6.40) 41.69 (5.84)	0.001
AFM (%)	Total Men Women	498 240 258	26.84 (11.82) 27.20 (12.49) 26.51 (11.17)	0.522
GFM (%)	Total Men Women	498 240 258	44.70 (18.64) 38.68 (14.44) 56.10 (18.20)	0.001

• BMI: Body Mass Index; WC: Waist Circumference; WHR: Waist-to-Hip Ratio; BF: Body Fat; AFM: Android Fat Mass; GFM: Gynoid Fat Mass; N: number of individuals; SD: Standard deviation.

Analysis of population stratification

When individuals in a population do not breed randomly, the population is said to be stratified. In recently admixed populations, the presence of stratification implies the existence of groups with differences in their genetic composition. These differences must be taken into account in association studies, as they can lead to spurious association results. To check if the population is under stratification, HWE tests were carried out for all the genetic markers studied. The observed and expected heterozygosity values and corresponding p-values are detailed in Supplementary Material - Table SI.

Fourteen out of the 46 AIM-InDels showed statistically significantly lower heterozygozities than the expected (p < 0.05): MID-881, MID-3122, MID-593, MID-1644, MID-3854, MID-3072, MID-2264, MID-419, MID-493, MID-159, MID-2005, MID-1802, MID-1386 and MID-3626.

Among the four SNPs, rs4074134 (BDNF) had its observed heterozygosity (0.309) statistically significantly lower (p = 0.001) than the expected (0.395).

These results show deviations to Hardy-Weinberg Equilibrium, indicating population stratification. The homozygous charge is greater than the expected if it was a non-stratified population. In association studies, when a population is under stratification, spurious associations between genetic factors and characteristics linked to diseases may arise from differences in ancestry that are not related to the risk of disease itself.

The use of AIMs as a correction factor in association analyses between SNPs and the obesity-related markers related to a clinical condition is vital so that these stratification effects do not bias the results and lead to misinterpretations.

Genomic ancestry profile of the studied sample

A cluster analysis was performed, to estimate the proportions of African, European and Native American ancestry in the 501 individuals included in our sample, using the software STRUCTURE. This analysis was based on the 46 AIM-InDel profiles.

The average ancestry values in our sample from Rio de Janeiro were 28.80% African, 57.20% European, and 14% Native American. The results obtained reveal, however, a large variation among individuals (Figure 1). The variation in the proportions of Native American contribution was relatively small (SD = 0.11, Variance = 0.01) when compared to the African (SD = 0.25, Variance = 0.06) and European (SD = 0.25, Variance = 0.06) ancestries.

Figure 1
Genomic ancestry estimates of each individual in the population sample of the State of Rio de Janeiro. The average genomic ancestry values in the sample were 28.8% African, 57.2% European, and 14% Native American. Each column represents an individual and each color represents the corresponding relative ancestry proportion in relation to a reference population. The analyses were carried out using reference population (option USEPOPINFO in STRUCTURE), and considering K = 3.

Obesity-related traits regarding the genomic ancestry of the sampled population

The quantitative characteristics of the obesity-related markers for each genetic ancestry of the population sample are summarized in Table II. Individuals with the majority of African ancestry exhibited the highest means for BMI, waist circumference, body fat percentage and gynoid fat mass, and the lowest for waist-to-hip ratio. The highest mean of android fat mass was detected in individuals with most European ancestry, while the highest for waist-to-hip ratio were observed in individuals with most European and most Native American ancestries alike. Individuals with most Native American ancestry showed the lowest means for all the obesity-related traits, with the exception of the waist-to-hip ratio, which there was no relevant difference between the three ancestries. There was no statistically significant difference observed.

Table II
Means, standard deviations and p-values of the obesity-related parameters for each genomic ancestry, as well as the frequencies of each ancestry in the population sample.

	AFR	EUR	NAM	p
n	123	359	16
BMI (kg/m²)	28.54 (5.11)	27.97 (5.06)	25.81 (3.80)	0.111
WC (cm)*	97.69 (12.12)	97.09 (12.65)	92.81 (10.66)	0.338
WHR*	0.93 (0.07)	0.94 (0.08)	0.94 (0.08)	0.805
BF (%)	37.63 (8.78)	35.88 (8.17)	34.15 (6.74)	0.077
AFM (%)	26.92 (11.42)	27.03 (12.05)	22.06 (8.88)	0.257
GFM (%)	50.46 (19.19)	47.07 (18.60)	40.44 (11.71)	0.062

• BMI: Body Mass Index; WC: Waist Circumference; WHR: Waist-to-Hip Ratio; BF: Body Fat; AFM: Android Fat Mass; GFM: Gynoid Fat Mass. *For WC and WHR n = 501: AFR = 125, EUR = 360, NAM = 16.

Association between genomic ancestry and the obesity-related markers

Although we did not observe statistically significant associations, it is worth emphasizing the inverse association values between Native American ancestry and waist circumference, body fat percentage, android fat mass and gynoid fat mass (Table III). In the analysis without the inclusion of correction variables, according to the unstandardized regression coefficient (β), we found a statistically significant direct association between African ancestry and body fat percentage (β = 3.170, p = 0.036) and an inverse association between Native American ancestry and gynoid fat mass (β = -16.330, p = 0.040).

Table III
Multiple linear regression association and correlation analyses results, as well as confidence intervals and p-values, between the genomic ancestries and the obesity-related parameters.

	β (IC 95 %)	r	p
	No control variables
BMI (kg/m²)
AFR	1.051 (-0.756, 2.859)	0.051	0.254
EUR	-0.597 (-2.386, 1.192)	-0.029	0.512
NAM	-2.417 (-6.664, 1.829)	-0.050	0.264
WC (cm)
AFR	-0.313 (-4.763, 4.137)	-0.006	0.890
EUR	1.468 (-2.930, 5.866)	0.029	0.512
NAM	-6.533 (-16.967, 3.900)	-0.055	0.219
WHR
AFR	-0.018 (-0.045, 0.008)	-0.061	0.176
EUR	0.018 (-0.008, 0.045)	0.061	0.172
NAM	-0.002 (-0.065, 0.061)	-0.003	0.954
BF (%)
AFR	3.170 (0.204, 6.137)	0.094	0.036
EUR	-1.881 (-4.817, 1.056)	-0.056	0.209
NAM	-6.813 (-13.772, 0.146)	-0.086	0.055
AFM (%)
AFR	-0.680 (-4.913, 3.553)	-0.014	0.752
EUR	2.306 (-1.869, 6.480)	0.049	0.278
NAM	-9.246 (-19.137, 0.646)	-0.082	0.067
GFM (%)
AFR	6.441 (-0.212, 13.095)	0.085	0.058
EUR	-3.383 (-9.968, 3.203)	-0.045	0.313
NAM	-16.330 (-31.920, -0.740)	-0.092	0.040
	With the control variables age, gender and income
BMI (kg/m²)
AFR	0.784 (-1.128, 2.696)	0.036	0.421
EUR	-0.264 (-2.135, 1.606)	-0.013	0.781
NAM	-2.518 (-6.786, 1.750)	-0.052	0.247
WC (cm)
AFR	0.202 (-4.476, 4.881)	0.004	0.932
EUR	1.198 (-3.376, 5.772)	0.023	0.607
NAM	-7.237 (-17.664, 3.191)	-0.061	0.173
WHR
AFR	-0.007 (-0.032, 0.017)	-0.027	0.553
EUR	0.009 (-0.015, 0.033)	0.033	0.469
NAM	-0.009 (-0.064, 0.046)	-0.015	0.746
BF (%)
AFR	-0.874 (-3.191, 1.444)	-0.033	0.459
EUR	1.517 (-0.743, 3.777)	0.059	0.188
NAM	-3.560 (-8.721, 1.602)	-0.061	0.176
AFM (%)
AFR	-0.428 (-4.924, 4.068)	-0.008	0.852
EUR	2.265 (-2.121, 6.651)	0.046	0.311
NAM	-9.670 (-19.662, 0.322)	-0.085	0.058
GFM (%)
AFR	1.277 (-4.973, 7.527)	0.018	0.688
EUR	1.292 (-4.812, 7.395)	0.019	0.678
NAM	-13.085 (-26.979, 0.809)	-0.083	0.065

• BMI: Body Mass Index; WC: Waist Circumference; WHR: Waist-to-Hip Ratio; BF: Body Fat; AFM: Android Fat Mass; GFM: Gynoid Fat Mass.

To avoid spurious interpretations, a new analysis was performed adjusting for age, gender and income variables. No statistically significant associations were detected. The new p-values were above the significance level between African ancestry and body fat percentage (β = -0.874, p = 0.459), and Native American ancestry and gynoid fat mass (β = -13.085, p = 0.065), which indicates an impact of the control variables in the unadjusted analysis.

It is worth noting that in our samples, as well as in a previously studied sample from the State of Rio de Janeiro population (Manta et al. 2013), most individuals have low Native American ancestry, and the variation in the African contribution is quite complementary to the European one. This may explain the opposite non-statistically significant associations observed between these two ancestries and the obesity-related markers.

Association between the SNPs of MC4R, FTO, FAIM2, BDNF and the obesity-related markers

We observed a statistically significant direct association between the SNP 4074134 (BDNF) CC genotype and gynoid fat mass (Table IV).

Table IV
Multiple linear regression association and correlation analyses results, as well as confidence intervals and p-values, between the SNP genotypes and the obesity-related parameters.

	β (IC 95 %)	r	p
No control variables
BMI (kg/m²)
rs17782313 CT	0.536 (-0.424, 1.496)	0.043	0.273
rs17782313 CC*	1.013 (-1.090, 3.117)	0.035	0.344
rs9939609 AT	-0.542 (-1.542, 0.457)	-0.066	0.287
rs9939609 AA*	0.354 (-0.956, 1.664)	0.051	0.596
rs7138803 AG	-0.538 (-1.468, 0.392)	-0.056	0.256
rs7138803 AA*	0.249 (-1.480, 1.978)	0.026	0.777
rs4074134 CT	-0.118 (-1.625, 1.389)	-0.091	0.878
rs4074134 CC*	1.068 (-0.335, 2.472)	0.113	0.135
WC (cm)
rs17782313 CT	1.106 (-1.256, 3.468)	0.034	0.358
rs17782313 CC*	2.678 (-2.511, 7.868)	0.039	0.311
rs9939609 AT	-0.587 (-3.048, 1.874)	-0.030	0.640
rs9939609 AA*	0.481 (-2.738, 3.700)	0.025	0.769
rs7138803 AG	-1.433 (-3.717, 0.852)	-0.064	0.218
rs7138803 AA*	1.371 (-2.890, 5.631)	0.043	0.528
rs4074134 CT	-0.299 (-4.023, 3.426)	-0.077	0.875
rs4074134 CC*	2.157 (-1.309, 5.623)	0.094	0.222
WHR
rs17782313 CT	-0.007 (-0.021, 0.008)	-0.043	0.363
rs17782313 CC*	0.003 (-0.028, 0.034)	0.015	0.847
rs9939609 AT	-0.005 (-0.020, 0.010)	-0.024	0.516
rs9939609 AA*	-0.004 (-0.023, 0.016)	-0.004	0.704
rs7138803 AG	-0.003 (-0.017, 0.010)	-0.035	0.640
rs7138803 AA*	0.015 (-0.010, 0.041)	0.059	0.244
rs4074134 CT	0.002 (-0.021, 0.024)	0.040	0.886
rs4074134 CC*	0.010 (-0.011, 0.031)	0.057	0.353
BF (%)
rs17782313 CT	1.405 (-0.169, 2.980)	0.070	0.080
rs17782313 CC*	2.321 (-1.134, 5.775)	0.047	0.187
rs9939609 AT	0.063 (-1.584, 1.710)	0.018	0.940
rs9939609 AA*	1.039 (-1.110, 3.188)	0.046	0.343
rs7138803 AG	-1.673 (-3.198, -0.147)	-0.093	0.032
rs7138803 AA*	-0.794 (-3.632, 2.044)	-0.001	0.583
rs4074134 CT	0.261 (-2.230, 2.752)	0.054	0.837
rs4074134 CC*	1.488 (-0.829, 3.805)	0.077	0.208
AFM (%)
rs17782313 CT	1.399 (-0.843, 3.641)	0.049	0.221
rs17782313 CC*	2.341 (-2.578, 7.260)	0.034	0.350
rs9939609 AT	-0.562 (-2.904, 1.779)	-0.032	0.637
rs9939609 AA*	0.562 (-2.493, 3.617)	0.029	0.718
rs7138803 AG	-1.388 (-3.562, 0.785)	-0.060	0.210
rs7138803 AA*	0.420 (-3.624, 4.463)	0.024	0.838
rs4074134 CT	0.415 (-3.110, 3.940)	0.085	0.817
rs4074134 CC*	3.130 (-0.150, 6.409)	0.119	0.061
GFM (%)
rs17782313 CT	3.501 (-0.028, 7.029)	0.081	0.052
rs17782313 CC*	4.090 (-3.652, 11.832)	0.033	0.300
rs9939609 AT	0.237 (-3.450, 3.923)	0.027	0.900
rs9939609 AA*	3.565 (-1.245, 8.374)	0.070	0.146
rs7138803 AG	-3.950 (-7.367, -0.533)	-0.102	0.024
rs7138803 AA*	-0.795 (-7.152, 5.561)	-0.014	0.806
rs4074134 CT	1.847 (-3.708, 7.401)	0.078	0.514
rs4074134 CC*	6.049 (0.883, 11.216)	0.125	0.022
With the control variables age, gender and income; African, European and Native American ancestries
BMI (kg/m²)
rs17782313 CT	0.530 (-0.435, 1.495)	0.049	0.281
rs17782313 CC*	0.816 (-1.302, 2.934)	0.034	0.449
rs9939609 AT	-0.551 (-1.553, 0.451)	-0.049	0.280
rs9939609 AA*	0.392 (-0.931, 1.714)	0.026	0.561
rs7138803 AG	-0.054 (-1.537, 1.398)	-0.054	0.235
rs7138803 AA*	0.254 (-1.492, 2.000)	0.013	0.775
rs4074134 CT	-0.145 (-1.659, 1.370)	-0.009	0.851
rs4074134 CC*	1.018 (-0.400, 2.435)	0.064	0.159
WC (cm)
rs17782313 CT	1.483 (-0.875, 3.841)	0.056	0.217
rs17782313 CC*	2.807 (-2.385, 8.000)	0.048	0.289
rs9939609 AT	-0.481 (-2.935, 1.973)	-0.017	0.700
rs9939609 AA*	0.963 (-2.269, 4.195)	0.026	0.558
rs7138803 AG	-0.080 (-1.356, 2.341)	-0.079	0.078
rs7138803 AA*	0.765 (-3.511, 5.042)	0.016	0.725
rs4074134 CT	-0.270 (-3.994, 3.453)	-0.006	0.887
rs4074134 CC*	2.149 (-1.333, 5.630)	0.055	0.226
WHR
rs17782313 CT	0.001 (-0.012, 0.013)	0.003	0.940
rs17782313 CC*	0.008 (-0.020, 0.036)	0.026	0.571
rs9939609 AT	-0.003 (-0.016, 0.010)	-0.019	0.678
rs9939609 AA*	0.003 (-0.014, 0.020)	0.016	0.724
rs7138803 AG	-0.081 (-0.097, 0.074)	-0.091	0.053
rs7138803 AA*	0.003 (-0.019, 0.026)	0.013	0.779
rs4074134 CT	0.003 (-0.017, 0.022)	0.012	0.792
rs4074134 CC*	0.009 (-0.010, 0.027)	0.042	0.347
BF (%)
rs17782313 CT	0.570 (-0.598, 1.738)	0.043	0.338
rs17782313 CC*	0.809 (-1.759, 3.377)	0.028	0.536
rs9939609 AT	0.113 (-1.103, 1.328)	0.008	0.856
rs9939609 AA*	0.809 (-0.789, 2.406)	0.045	0.320
rs7138803 AG	-0.014 (-1.115, 1.258)	-0.019	0.673
rs7138803 AA*	0.343 (-1.775, 2.461)	0.014	0.750
rs4074134 CT	0.287 (-1.551, 2.124)	0.014	0.759
rs4074134 CC*	1.506 (-0.212, 3.224)	0.078	0.086
AFM (%)
rs17782313 CT	1.536 (-0.724, 3.796)	0.060	0.182
rs17782313 CC*	2.400 (-2.568, 7.369)	0.043	0.343
rs9939609 AT	-0.540 (-2.895, 1.815)	-0.020	0.653
rs9939609 AA*	0.871 (-2.224, 3.966)	0.025	0.581
rs7138803 AG	-0.066 (-1.356, 2.457)	-0.065	0.148
rs7138803 AA*	0.226 (-3.869, 4.322)	0.005	0.914
rs4074134 CT	0.355 (-3.198, 3.908)	0.009	0.844
rs4074134 CC*	3.179 (-0.143, 6.500)	0.085	0.061
GFM (%)
rs17782313 CT	1.938 (-1.208, 5.084)	0.055	0.227
rs17782313 CC*	1.764 (-5.152, 8.681)	0.023	0.616
rs9939609 AT	0.101 (-3.171, 3.373)	0.003	0.952
rs9939609 AA*	2.860 (-1.440, 7.160)	0.059	0.192
rs7138803 AG	-0.051 (-1.264, 3.021)	-0.057	0.205
rs7138803 AA*	1.456 (-4.242, 7.154)	0.023	0.616
rs4074134 CT	1.721 (-3.199, 6.642)	0.031	0.492
rs4074134 CC*	6.232 (1.632, 10.832)	0.120	0.008

• BMI: Body Mass Index; WC: Waist Circumference; WHR: Waist-to-Hip Ratio; BF: Body Fat; AFM: Android Fat Mass; GFM: Gynoid Fat Mass. rs17782313 TT, rs9939609 TT, rs7138803 GG and rs4074134 TT are the lowest risk genotypes described in association with obesity and metabolic conditions (Ensembl Genome Browser 2021), and were used as reference for the other two genotypes in the analyses and interpretations. *Highest risk genotypes.

In the analysis without correction variables, the unstandardized regression coefficient (β) denoted statistically significant inverse associations between the SNP rs7138803 (FAIM2) AG genotype, body fat percentage (β = -1.673, p = 0.032) and gynoid fat mass (β = -3.950, p = 0.024); and a statistically significant direct association between the SNP 4074134 (BDNF) CC genotype and gynoid fat mass (β = 6.049, p = 0.022).

A new analysis after adjusting for the variation in ancestry, age, sex and income was made and only the SNP 4074134 (BDNF) CC genotype direct association with gynoid fat mass persisted at the significance level (β = 6.232, p = 0.008), which points to an influence of the correction variables in the unadjusted association analysis between the SNP rs7138803 (FAIM2) AG genotype, body fat percentage (β = -0.014, p = 0.673) and gynoid fat mass (β = -0.051, p = 0.205). The SNP 4074134 (BDNF) CC genotype showed a significant higher percentage (6.23%) of gynoid fat mass compared to the lowest risk rs4074134 (BDNF) TT genotype.

Mean values of the obesity-related traits for each SNP genotype

Although there was no statistically significant difference noticed, we observed that individuals with the SNP highest risk genotypes have higher means of the obesity-related markers, with the exception of individuals with the highest risk genotypes of rs9939609 (FTO) for waist-to-hip ratio and rs7138803 (FAIM2) for body fat percentage and gynoid fat mass markers (Table V). This denotes, regardless of statistical significance, which depends on the sample size, some weight of the highest risk genotypes in the susceptibility of higher anthropometric and body composition measures, and consequently, adiposity.

Table V
Means, standard deviations and p-values of the obesity-related parameters for each genotype of the four genetic loci, as well as the frequencies of each genotype in the population sample.

	n	BMI (kg/m²)	WC (cm)**	WHR**	BF (%)	AFM (%)	GFM (%)
rs17782313 TT	311	27.80 (5.20)	96.57 (12.67)	0.94 (0.08)	35.68 (8.46)	26.21 (11.81)	46.35 (19.47)
rs17782313 CT	163	28.36 (4.79)	97.72 (12.11)	0.93 (0.08)	37.09 (8.02)	27.67 (11.66)	49.86 (16.98)
rs17782313 CC*	24	28.83 (4.67)	99.29 (12.16)	0.94 (0.07)	38.00 (8.32)	28.61 (12.73)	50.44 (18.06)
rs9939609 TT	164	28.21 (5.32)	97.23 (13.26)	0.94 (0.07)	36.03 (8.32)	26.92 (12.70)	46.94 (17.91)
rs9939609 AT	246	27.71 (4.87)	96.72 (11.77)	0.93 (0.08)	36.10 (8.46)	26.45 (11.20)	47.19 (19.34)
rs9939609 AA*	88	28.60 (5.02)	97.79 (12.89)	0.93 (0.08)	37.08 (8.00)	27.58 (11.80)	50.52 (18.07)
rs7138803 GG	251	28.22 (5.18)	97.55 (12.70	0.93 (0.08)	37.02 (8.27)	27.33 (11.83)	49.44 (18.15)
rs7138803 AG	209	27.71 (4.78)	96.17 (12.14)	0.93 (0.07)	35.35 (8.44)	26.01 (11.47)	45.48 (17.22)
rs7138803 AA*	38	28.50 (5.61)	98.97 (12.63)	0.95 (0.07)	36.23 (7.68)	27.82 (13.41)	48.64 (27.12)
rs4074134 TT	59	27.34 (4.78)	95.71 (11.87)	0.93 (0.07)	35.28 (9.38)	24.60 (11.26)	43.58 (19.14)
rs4074134 CT	152	27.35 (4.57)	95.65 (11.76)	0.93 (0.08)	35.58 (8.39)	25.33 (10.80)	45.51 (16.22)
rs4074134 CC*	287	28.53 (5.30)	98.11 (12.87)	0.94 (0.07)	36.81 (8.04)	28.04 (12.30)	49.71 (19.53)
p		0.511	0.193	0.092	0.374	0.453	0.492

• BMI: Body Mass Index; WC: Waist Circumference; WHR: Waist-to-Hip Ratio; BF: Body Fat; AFM: Android Fat Mass; GFM: Gynoid Fat Mass; SD: Standard Deviation. *Highest risk genotypes. **For WC and WHR n = 501: rs17782313 TT = 313, rs17782313 CT = 164, rs17782313 CC = 24; rs9939609 TT = 166, rs9939609 AT = 246, rs9939609 AA = 89; rs7138803 GG = 252, rs7138803 AG = 211, rs7138803 AA = 38; rs4074134 TT = 59, rs4074134 CT = 152, rs4074134 CC = 290.

DISCUSSION

Ancestry informative markers

The European and African demographic impact in Brazil is historically well documented and supported by genetic studies. Uniparental markers showed clear evidence of a biased admixture involving predominantly descendants of European men and African women (Alves-Silva et al. 2000, Carvalho-Silva et al. 2001, Hünemeier et al. 2007).

Studies using autosomal AIMs have demonstrated an essentially tri-hybrid heritage and a predominantly European ancestry in most Brazilian populations, with differences in ancestry proportions depending on the geographic region considered or in different self-declared ethnic groups (de Souza et al. 2019, Pena et al. 2020).

The results from this work are also in line with those from Manta et al. (2013). Through the analysis of 46 AIM-InDels in a population sample from Rio de Janeiro, they found a predominantly European ancestry (55.20%), followed by African (31.10%) and Amerindian (13.70%).

Considering the results described above, it is possible to conclude that the population ancestry estimates obtained are in accordance with previous studies, as well as with historical data concerning the origin of the Brazilian population.

Obesity-related traits

The obesity-related measures used in this study are good indicators of nutritional status and predictors of overweight and obesity. The BMI, waist circumference, waist-to-hip ratio and body fat percentage are the most used parameters in epidemiological studies. The present study has both strengths and limitations. Some of its strengths are its study design, ancestry estimation and anthropometric and body composition assessment. The latter includes the evaluation of body fat, android and gynoid fat masses as measured via DXA, therefore overcoming the limitations of previous studies that used only reported weights and anthropometric measurements such as BMI, WC, and WHR. The DXA accurately reflects the measures of body composition. This method allows accurate quantifications of total body fat as well as its distribution and location. The DXA is considered the gold standard in validation studies of methods and equations, and it is noninvasive, precise, and automatic (da Rocha et al. 2017).

Some study limitations should be mentioned. The sample size and the absence of information on physical activity, dietary habits and other possible residual confounding, which are recognized to influence our main outcome variables, may have limited our conclusions. Being a cross-sectional study, another limitation is the inability to establish temporality.

Genomic ancestry and the obesity-related traits

In general, studies examining associations between African or European ancestry using AIMs and markers related to obesity such as BMI report uniform findings. Direct associations are observed between African ancestry and traits such as BMI and waist circumference, while inverse relationships are detected with European ancestry. Inverse associations were also observed with Native American ancestry. However, the authors emphasize the important role of the environment and socioeconomic components in these results (Fernández et al. 2003, Tang et al. 2006, Lai et al. 2009, Cheng et al. 2010, Lins et al. 2012, Nassir et al. 2012, Goonesekera et al. 2015, Klimentidis et al. 2016).

In the study by Nassir et al. (2012) in post-menopausal African-American and Hispanic-American participants of the Women’s Health Initiative (WHI), significant direct associations were found between African ancestry and BMI in the general population, as well as in subgroups of African-Americans and Hispanic-Americans. However, when waist-to-hip ratio was used as a parameter of obesity, a significantly weaker direct association with African ancestry was found in the general population and unobserved among African-Americans or Hispanic-Americans. Interestingly, Amerindian ancestry was significantly direct associated with waist-to-hip ratio, but not with BMI in the general population.

Amerindian ancestry showed non-significant inverse associations for all obesity-related markers used in this study. One explanation may be the diet and healthy lifestyle. Goonesekera et al. (2015) observed evidence of a modifying effect on the associations between Native American ancestry and adiposity by diet. Individuals with this ancestry who had a relatively high healthy eating score, had inverse associations with all anthropometric and body composition measures, although not all results have reached statistical significance.

We did not replicate the previous findings of a statistically significant association between genomic ancestry, anthropometric parameters and body composition measures. The differences in environmental, social, economic and cultural conditions between the populations investigated in the different studies may explain our lack of significant association. However, our findings must be further replicated in larger samples using populations from the same origin.

When studying characteristics related to health, biodiversity and variability attributed to the use of genetic admixture estimates, they should not be simply replaced by the simple classification of ethnicity, but used together. Self-declaration of ethnicity is easier to access and important to personal identification, but its exclusive use confuses the historical ancestry of individuals and the understanding of the biological factors underlying obesity, its traits related to metabolic risks, and its variation within and between populations (Fernández et al. 2003).

As populations continue to breed, self-identification tends to become less a reflection of an individual true ancestral origin. Divers et al. (2011) described this trend in which, among Hispanic Americans in the Multi-Ethnic Study for Atherosclerosis (MESA), genomic ancestry did not correspond to Hispanic self-declaration in 30% of its self-declared Hispanic participants.

Relationship between the SNPs and the obesity-related traits

rs17782313 (MC4R)

A possible explanation for not finding a statistically significant association between this locus and obesity-related markers, in this study, is that it may be more associated with obesity in childhood than in adulthood (Hardy et al. 2009). The rs17782313 showed stronger direct associations with total mass than with BMI. Its association with mass increased during childhood and adolescence, peaked at the age of 20, and weakened during adulthood. It was concluded that the genetic variants in MC4R showed similar biphasic changes in their associations with BMI and weight, respectively, strengthening during childhood until the age of 20 and weakening with increasing adulthood.

We failed to identify a significant association between the SNP and the obesity-related measures in a Brazilian population sample. Da Fonseca et al. (2019) observed a similar result. They found no significant association between the SNP and obesity in a cohort of 490 individuals from Brazil. Grant et al. (2009) also found a similar result, in which no significant association was observed between rs17782313 (MC4R) and obesity in Afro-American individuals. These findings suggest that the influence of this polymorphism on the obesity risk may also be dependent on the ethnic group, and on different environmental, social and cultural characteristics.

rs9939609 (FTO)

In a recent study in which male carriers of the risk allele (A) were investigated, they had body mass, BMI, WHR and waist and hip circumferences significantly higher than men without this risk allele. They were also more frequently diagnosed with obesity based on BMI and waist circumference. This influence was not seen in women. The results point to a male-specific association between the FTO polymorphism and obesity characteristics. (Zdrojowy-Wełna et al. 2020). Manco et al. (2019), in turn, observed an direct association between rs9939609 (FTO) and obesity or body fat indices in female adolescents, but not in males.

The role of FTO variants on obesity risk is consistent in studies with European ancestry individuals (Dina et al. 2007). However, apparently it is not associated with obesity risk in African Americans, Chinese Han, and Native Oceanic populations (Ohashi et al. 2007, Scuteri et al. 2007, Li et al. 2008).

In our investigation, we did not replicate the previous findings that suggest significant association between the polymorphism and the obesity-related parameters. Other studies have analysed the relationship between rs9939609 (FTO) and obesity susceptibility in Brazil. Reuter et al. (2016) studied the association between this SNP and overweight/obesity risk in a sample of youths from the South of Brazil. They observed that the rs9939609 (FTO) was significantly associated with increased BMI and waist circumference. Pereira et al. (2016) have performed a similar study with children. However, they did not find this association.

Ramos et al. (2012) investigated the contribution of rs9939609 (FTO) in morbid obesity risk. Their results suggested that the presence of the rs9939609 polymorphism risk allele increased the risk for obesity. More recently, Da Fonseca et al. (2019, 2020) observed that this polymorphism plays a significant role in predisposing higher body weight, BMI and severe obesity in a Brazilian population sample. The conflicting results may be explained by the differences in social and cultural conditions of the individuals, such as income and eating habits.

rs7138803 (FAIM2)

Our study is the first in a Brazilian population to investigate the association of rs7138803 (FAIM2) with obesity-related adiposity markers. Although we did not observe statistically significant association in our population sample, other authors show an association of this polymorphism with obesity-related parameters, mainly in European and Asian populations. These contrasting findings are probably due to significant differences in ethnic, environmental, social and cultural components along the continents.

Thorleifsson et al. (2009) found that rs7138803 (FAIM2) was associated with increased BMI and obesity based on studies of GWAS in individuals of European ancestry. Subsequently, a replicated study showed that this SNP was significantly associated with higher BMI, waist circumference and obesity risk in the Chinese population (Wu et al. 2010).

Recently, in Chinese Han population, Kang et al. (2020) observed associations with increased BMI, diastolic blood pressure and triglycerides. However, in the same Han population, Li et al. (2013) detected no significant effect of rs7138803 (FAIM2) in obesity. Another study showed that the polymorphism was associated in adiposity regulation in both Chinese Han individuals and Europeans (Ng et al. 2010). In Europeans, the rs7138803 (FAIM2) seems to contribute to variation in body morphology and composition in Spanish people. (Poveda et al. 2014).

rs4074134 (BDNF)

Xu et al. (2003) observed that, similar to the MC4R mutations, mice with mutations that express decreased amounts of the BDNF receptor showed hyperphagia and obesity at the beginning of maturity, suggesting a role of the gene in the energy balance.

In the present study, we observed a significant association between the rs4074134 (BDNF) and gynoid fat mass. Gynoid fat accumulation is often associated with a protective effect on cardiovascular risk, as opposed to android fat. Accumulation in gynoid fat is associated with a decreased risk of type 2 diabetes mellitus and cardiovascular risk, which in part may reflect the downregulation of proinflammatory signaling pathways and insulin resistance. The mechanisms for the cardiometabolic benefits of gynoid fat relate to its inverse association with insulin resistance and include less expression of proinflammatory cytokines compared with android fat mass (Min & Min 2015, Miller et al. 2020).

In contrast, Wiklund et al. (2008) observed that gynoid fat mass was significantly associated with impaired glucose tolerance, hypertriglyceridemia and hypertension and consequently cardiovascular risk, mainly in men. The inconsistency of the results may rely on possible confounding factors not adjusted.

Since most of the studies investigating the relationship between rs4074134 (BDNF) and obesity have been carried out in Asian populations, there is still much to be clarified in populations from other continents. Assessing the influence of fat distribution, and gynoid fat mass in particular, on cardiovascular risk endpoints such as stroke and heart diseases requires further investigation.

CONCLUSIONS

Our results suggest that the rs4074134 (BDNF) CC genotype was directly associated with gynoid fat mass in the population sample, independently of age, gender, income and genomic ancestry of the individuals. Statistically significant association was restricted to gynoid fat mass evaluated via DXA, with no evidence of association with the other body composition measures neither with the measured anthropometric indices. Given the high validity of our outcome measurements, this study findings strengthen available evidence about genetic factors on fat accumulation. We did not observe statistically significant associations between genomic ancestry and the markers related to obesity. We highlight, although not statistically significant, the results values of inverse association between Native American ancestry and the obesity-related traits in the sampled population. We strongly recommend association studies between genetic factors and obesity-related measures with larger sample sizes in the Brazilian population.

ACKNOWLEDGMENTS

The authors acknowledge support from the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and from the Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ). This study was financed in part by CAPES, Finance Code 001, and has been supported by the Brazilian funding agency FAPERJ, grant no.: E-26/010.0012661/2016. The authors report no conflict of interest.

SUPPLEMENTARY MATERIAL

Figure 1. Genomic ancestry estimates of each individual in the population sample of the State of Rio de Janeiro. The average genomic ancestry values in the sample were 28.8% African, 57.2% European, and 14% Native American. Each column represents an individual and each color represents the corresponding relative ancestry proportion in relation to a reference population. The analyses were carried out using reference population (option USEPOPINFO in STRUCTURE), and considering K = 3.

REFERENCES

• ALVES-SILVA J, DA SILVA SANTOS M, GUIMARÃES PEM, FERREIRA ACS, BANDELT HJ, PENA SDJ & PRADO VF. 2000. The ancestry of Brazilian mtDNA lineages. Am J Hum Genet 67(2): 444-461.
• CARVALHO-SILVA DR, SANTOS FR, ROCHA J & PENA SDJ. 2001. The phylogeography of Brazilian Y-chromosome lineages. Am J Hum Genet 68(1): 281-286.
• CHENG CY ET AL. 2010. Admixture mapping of obesity-related traits in african americans: The atherosclerosis risk in communities (ARIC) study. Obesity 18(3): 563-572.
• CHURCH C ET AL. 2010. Overexpression of Fto leads to increased food intake and results in obesity. Nat Genet 42(12): 1086-1092.
• DA FONSECA ACP, ABREU GM, ZEMBRZUSKI VM, CAMPOS JUNIOR M, CARNEIRO JRI, NETO JFN, CABELLO GMK & CABELLO PH. 2019. The association of the fat mass and obesity-associated gene (FTO) rs9939609 polymorphism and the severe obesity in a Brazilian population. Diabetes Metab Syndr Obes Targets Ther 12: 667-684.
• DA FONSECA ACP, MARCHESINI B, ZEMBRZUSKI VM, VOIGT DD, RAMOS VG, CARNEIRO JRI, NETO JFN, DE CABELLO GMK & CABELLO PH. 2020. Genetic variants in the fat mass and obesity-associated (FTO) gene confer risk for extreme obesity and modulate adiposity in a brazilian population. Genet Mol Biol 43(1): e20180264.
• DA FONSECA ACP, MASTRONARDI C, JOHAR A, ARCOS-BURGOS M & PAZ-FILHO G. 2017. Genetics of non-syndromic childhood obesity and the use of high-throughput DNA sequencing technologies. J Diabetes Complications 31(10): 1549-1561.
• DA ROCHA TF, HASSELMANN MH, CHAVES CURIONI C, BEZERRA FF & FAERSTEIN E. 2017. Alcohol consumption is associated with DXA measurement of adiposity: the Pró-Saúde Study, Brazil. Eur J Nutr 56(5): 1983-1991.
• DE SOUZA AM, RESENDE SS, DE SOUSA TN & DE BRITO CFA. 2019. A systematic scoping review of the genetic ancestry of the brazilian population. Genet Mol Biol 42(3): 495-508.
• DINA C et al. 2007. Variation in FTO contributes to childhood obesity and severe adult obesity. Nat Genet 39(6): 724-6.
• DIVERS J, REDDEN DT, RICE KM, VAUGHAN LK, PADILLA MA, ALLISON DB, BLUEMKE DA, YOUNG HJ & ARNETT DK. 2011. Comparing self-reported ethnicity to genetic background measures in the context of the Multi-Ethnic Study of Atherosclerosis (MESA). BMC Genet 12: 28.
• DORAJOO R ET AL. 2012. Replication of 13 obesity loci among Singaporean Chinese, Malay and Asian-Indian populations. Int J Obes 36: 159-163.
• DUCLOS M. 2016. Osteoarthritis, obesity and type 2 diabetes: The weight of waist circumference. Ann Phys Rehabil Med 59(3): 157-160.
• EXCOFFIER L & LISCHER HEL. 2010. Arlequin suite ver 3.5: A new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour 10(3): 564-567.
• FAERSTEIN E, CHOR D, LOPES CS & WERNECK GL. 2005. Estudo Pró-Saúde: características gerais e aspectos metodológicos. Rev Bras Epidemiol 8(4): 454-466.
• FERNÁNDEZ JR, PEARSON KE, KELL KP & BOHAN BROWN MM. 2013. Genetic admixture and obesity: Recent perspectives and future applications. Hum Hered 75(2-4): 98-105.
• FERNÁNDEZ JR ET AL. 2003. Association of African genetic admixture with resting metabolic rate and obesity among women. Obes Res 11(7): 904-911.
• FINUCANE MM ET AL. 2011. National, regional, and global trends in body-mass index since 1980: Systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9·1 million participants. Lancet 377(9765): 557-567.
• FRAYLING TM ET AL. 2007. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 316(5826): 889-894.
• GOONESEKERA SD, FANG SC, PICCOLO RS, FLOREZ JC & MCKINLAY JB. 2015. Biogeographic ancestry is associated with higher total body adiposity among African-American Females: The Boston Area Community Health Survey. PLoS One 10(4): e0122808.
• GRANT SFA ET AL. 2009. Investigation of the locus near MC4R with childhood obesity in Americans of European and African ancestry. Obesity 17(7): 1461-1465.
• GRANT SFA ET AL. 2008. Association analysis of the FTO gene with obesity in children of Caucasian and African ancestry reveals a common tagging SNP. PLoS One 3(3): e1746.
• HAN X, LUO Y, ZHANG X, LV C, SUN X, ZHANG X, ZHOU X, CAI X, REN Q & JI L. 2013. Rs4074134 Near BDNF Gene Is Associated with Type 2 Diabetes Mellitus in Chinese Han Population Independently of Body Mass Index. PLoS One 8(2): e56898.
• HARDY R, WILLS AK, WONG A, ELKS CE, WAREHAM NJ, LOOS RJF, KUH D & ONG KK. 2009. Life course variations in the associations between FTO and MC4R gene variants and body size. Hum Mol Genet 19(3): 545-552.
• HOTTA K ET AL. 2009. Association between obesity and polymorphisms in SEC16B, TMEM18, GNPDA2, BDNF, FAIM2 and MC4R in a Japanese population. J Hum Genet 54(12): 727-731.
• HOWE KL ET AL. 2021. Ensembl 2021. Nucleic Acids Res. 49(D1): D884-D891.
• HÜNEMEIER T, CARVALHO C, MARRERO AR, SALZANO FM, PENA SDJ & BORTOLINI MC. 2007. Niger-Congo speaking populations and the formation of the Brazilian gene pool: MtDNA and Y-chromosome data. Am J Phys Anthropol 133(2): 854-867.
• HUSZAR D ET AL. 1997. Targeted disruption of the melanocortin-4 receptor results in obesity in mice. Cell 88(1): 131-141.
• JAKOBSSON M & ROSENBERG NA. 2007. CLUMPP: A cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23(14): 1801-1806.
• KANG J, GUAN RC, ZHAO Y & CHEN Y. 2020. Obesity-related loci in TMEM18, CDKAL1 and FAIM2 are associated with obesity and type 2 diabetes in Chinese Han patients. BMC Med Genet 21(1): 65.
• KHALILITEHRANI A, QORBANI M, HOSSEINI S & PISHVA H. 2015. The association of MC4R rs17782313 polymorphism with dietary intake in Iranian adults. Gene 563(2): 125-129.
• KLIMENTIDIS YC, ARORA A, ZHOU J, KITTLES R & ALLISON DB. 2016. The genetic contribution of West-African ancestry to protection against central obesity in African-American men but not women: Results from the ARIC and MESA studies. Front Genet 7: 89.
• LAI CQ, TUCKER KL, CHOUDHRY S, PARNELL LD, MATTEI J, GARCÍA-BAILO B, BECKMAN K, BURCHARD EG & ORDOVÁS JM. 2009. Population admixture associated with disease prevalence in the Boston Puerto Rican health study. Hum Genet 125(2): 199-209.
• LI C, QIU X, YANG N, GAO J, RONG Y, XIONG C & ZHENG F. 2013. Common rs7138803 variant of FAIM2 and obesity in Han Chinese. BMC Cardiovasc Disord 13: 56.
• LI H, WU Y, LOOS RJF, HU FB, LIU Y, WANG J, YU Z & LIN X. 2008. Variants in the fat mass- and obesity-associated (FTO) gene are not associated with obesity in a Chinese Han population. Diabetes 57(1): 264-268.
• LINS TC, PIRES AS, PAULA RS, MORAES CF, VIEIRA RG, VIANNA LG, NOBREGA OT & PEREIRA RW. 2012. Association of serum lipid components and obesity with genetic ancestry in an admixed population of elderly women. Genet Mol Biol 35(3): 575-582.
• LOHMAN TJ, ROACHE AF & MARTORELL R. 1992. Anthropometric Standardization Reference Manual. Med Sci Sport Exerc 24(8): 952 p.
• LOOS RJF. 2011. The genetic epidemiology of melanocortin 4 receptor variants. Eur J Pharmacol 660(1): 156-164.
• LOOS RJF & BOUCHARD C. 2003. Obesity - Is it a genetic disorder? J Intern Med 254(5): 401-425.
• MANCO L, PINHO S, ALBUQUERQUE D, MACHADO-RODRIGUES AM & PADEZ C. 2019. Physical activity and the association between the FTO rs9939609 polymorphism and obesity in Portuguese children aged 3 to 11 years. Am J Hum Biol 31(6): e23312.
• MANTA FSN, PEREIRA R, CAIAFA A, SILVA DA, GUSMÃO L & CARVALHO EF. 2013. Analysis of genetic ancestry in the admixed Brazilian population from Rio de Janeiro using 46 autosomal ancestry-informative indel markers. Ann Hum Biol 40(1): 94-98.
• MILLER M, RYAN A, REED RM, GOGGINS C, SORKIN J & GOLDBERG AP. 2020. Effect of Icosapent Ethyl on Gynoid Fat and Bone Mineral Health in the Metabolic Syndrome: A Preliminary Report. Clin Ther 42(11): 2226-2230.
• MIN KB & MIN JY. 2015. Android and gynoid fat percentages and serum lipid levels in United States adults. Clin Endocrinol (Oxf) 82(3): 377-387.
• MINISTÉRIO DA SAÚDE. 2020. Vigilância de Fatores de Risco e Proteção para Doenças Crônicas por Inquérito Telefônico - VIGITEL 2019 [Surveillance of Risk and Protection Factors for Chronic Diseases by Telephone Survey - VIGITEL 2019]; [accessed 2021 Aug 13]. https://bvsms.saude.gov.br/bvs/publicacoes/vigitel_brasil_2019_vigilancia_fatores_risc o.pdf Portuguese.
  » https://bvsms.saude.gov.br/bvs/publicacoes/vigitel_brasil_2019_vigilancia_fatores_risc o.pdf
• NASSIR R ET AL. 2012. Relationship between adiposity and admixture in African-American and Hispanic-American women. Int J Obes 36(2): 304-313.
• NG M ET AL. 2014. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: A systematic analysis for the Global Burden of Disease Study 2013. Lancet 384(9945): 766-781.
• NG MCY, TAM CHT, SO WY, HO JSK, CHAN AW, LEE HM, WANG Y, LAM VKL, CHAN JCN & MA RCW. 2010. Implication of genetic variants near NEGR1, SEC16B, TMEM18, ETV5/DGKG, GNPDA2, LIN7C/BDNF, MTCH2, BCDIN3D/FAIM2, SH2B1, FTO, MC4R, and KCTD15 with obesity and type 2 diabetes in 7705 Chinese. J Clin Endocrinol Metab 95(5): 2418-2425.
• NINGOMBAM SS, CHHUNGI V, NEWMEI MK, RAJKUMARI S, DEVI NK, MONDAL PR & SARASWATHY KN. 2018. Differential distribution and association of FTO rs9939609 gene polymorphism with obesity: A cross-sectional study among two tribal populations of India with East-Asian ancestry. Gene 647: 198-204.
• OHASHI J ET AL. 2007. FTO polymorphisms in oceanic populations. J Hum Genet 52(12): 1031-1035.
• PATNAIK L, PATTNAIK S, RAO EV & SAHU T. 2017. Validating neck circumference and waist circumference as anthropometric measures of overweight/obesity in adolescents. Indian Pediatr 54(5): 377-380.
• PATRON J, SERRA-CAYUELA A, HAN B, LI C & WISHART DS. 2019. Assessing the performance of genome-wide association studies for predicting disease risk. PLoS One 14(12): e0220215.
• PENA SDJ, SANTOS FR & TARAZONA-SANTOS E. 2020. Genetic admixture in Brazil. Am J Med Genet Part C Semin Med Genet 184(4): 928-938.
• PEREIRA PA, ALVIM-SOARES AM, SANDRIM VC, LANNA CMM, SOUZA-COSTA DC, BELO VA, DE PAULA JJ, TANUS-SANTOS JE, ROMANO-SILVA MA & DE MIRANDA DM. 2016. Lack of association between genetic polymorphism of FTO, AKT1 and AKTIP in childhood overweight and obesity. J Pediatr 92(5): 521-527.
• PEREIRA R, PHILLIPS C, PINTO N, SANTOS C, DOS SANTOS SEB, AMORIM A, CARRACEDO Á & GUSMÃO L. 2012. Straightforward inference of ancestry and admixture proportions through ancestry-informative insertion deletion multiplexing. PLoS One 7(1): e29684.
• PIGEYRE M, YAZDI FT, KAUR Y & MEYRE D. 2016. Recent progress in genetics, epigenetics and metagenomics unveils the pathophysiology of human obesity. Clin Sci 130(12): 943-986.
• POVEDA A, IBÁÑEZ ME & REBATO E. 2014. Common variants in BDNF, FAIM2, FTO, MC4R, NEGR1, and SH2B1 show association with obesity-related variables in Spanish Roma population. Am J Hum Biol 26(5): 660-669.
• PRITCHARD JK, STEPHENS M & DONNELLY P. 2000. Inference of population structure using multilocus genotype data. Genetics 155(2): 945-59.
• QASIM A, TURCOTTE M, DE SOUZA RJ, SAMAAN MC, CHAMPREDON D, DUSHOFF J, SPEAKMAN JR & MEYRE D. 2018. On the origin of obesity: identifying the biological, environmental and cultural drivers of genetic risk among human populations. Obes Rev 19(2): 121-149.
• RAMOS AV, BASTOS-RODRIGUES L, RESENDE BA, FRIEDMAN E, CAMPANHA-VERSIANI L, MIRANDA DM, SARQUIS M & DE MARCO L. 2012. The contribution of FTO and UCP-1 SNPs to extreme obesity, diabetes and cardiovascular risk in Brazilian individuals. BMC Med Genet 13: 101.
• REUTER CP ET AL. 2016. FTO polymorphism, cardiorespiratory fitness, and obesity in Brazilian youth. Am J Hum Biol 28(3): 381-386.
• SABIN MA, WERTHER GA & KIESS W. 2011. Genetics of obesity and overgrowth syndromes. Best Pract Res Clin Endocrinol Metab 25(1): 207-220.
• SCUTERI A et al. 2007. Genome-wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits. PLoS Genet 3(7): e115.
• SHABANA & HASNAIN S. 2015. Effect of the common fat mass and obesity associated gene variants on obesity in Pakistani population: A case-control study. Biomed Res Int 2015: 852920.
• SHRIVER MD, SMITH MW, JIN L, MARCINI A, AKEY JM, DEKA R & FERRELL RE. 1997. Ethnic-affiliation estimation by use of population-specific DNA markers. Am J Hum Genet 60(4): 957-964.
• TANG H, JORGENSON E, GADDE M, KARDIA SLR, RAO DC, ZHU X, SCHORK NJ, HANIS CL & RISCH N. 2006. Racial admixture and its impact on BMI and blood pressure in African and Mexican Americans. Hum Genet 119(6): 624-633.
• THORLEIFSSON G et al. 2009. Genome-wide association yields new sequence variants at seven loci that associate with measures of obesity. Nat Genet 41(1): 18-24.
• TSCHRITTER O, HAUPT A, PREISSL H, KETTERER C, HENNIGE AM, SARTORIUS T, MACHICAO F, FRITSCHE A & HÄRING HU. 2011. An obesity risk SNP (rs17782313) near the MC4R gene is associated with cerebrocortical insulin resistance in humans. J Obes 2011: 283153.
• VEGA JA, SALAZAR G, ISABEL HODGSON M, CATALDO LR, VALLADARES M, OBREGÓN AM & SANTOS JL. 2016. Melanocortin-4 Receptor Gene Variation Is Associated with Eating Behavior in Chilean Adults. Ann Nutr Metab 68(1): 35-41.
• VELÁZQUEZ-FERNÁNDEZ D ET AL. 2017. Analysis of Gene Candidate SNP and Ancestral Origin Associated to Obesity and Postoperative Weight Loss in a Cohort of Obese Patients Undergoing RYGB. Obes Surg 27(6): 1481-1492.
• WAALEN J. 2014. The genetics of human obesity. Transl Res 164(4): 293-301.
• WELLS JCK. 2006. The evolution of human fatness and susceptibility to obesity: An ethological approach. Biol Rev Camb Philos Soc 81(2): 183-205.
• WIKLUND P, TOSS F, WEINEHALL L, HALLMANS G, FRANKS PW, NORDSTRÖM A & NORDSTRÖM P. 2008. Abdominal and gynoid fat mass are associated with cardiovascular risk factors in men and women. J Clin Endocrinol Metab 93(11): 4360-4366.
• WORLD HEALTH ORGANIZATION. 2008. Waist Circumference and Waist-Hip Ratio: Report of a WHO Expert Consultation, 39 p.
• WORLD HEALTH ORGANIZATION. 2020. Obesity and overweight; [accessed 2021 Aug 14]. https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight
  » https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight
• WORLD MEDICAL ASSOCIATION. 2013. Declaration of Helsinki. Bulletin World Health Organ; [accessed 2021 Aug 14]. https://www.who.int/bulletin/archives/79(4)373.pdf
  » https://www.who.int/bulletin/archives/79(4)373.pdf
• WU L ET AL. 2010. Associations of six single nucleotide polymorphisms in obesity-related genes with BMI and risk of obesity in chinese children. Diabetes 59(12): 3085-3089.
• WU L, ZHAO X, SHEN Y, ZHANG MX, YAN Y, HOU D, MENG L, LIU J, CHENG H & MI J. 2015. Promoter methylation of fas apoptotic inhibitory molecule 2 gene is associated with obesity and dyslipidaemia in Chinese children. Diabetes Vasc Dis Res 12(3): 217-220.
• WU Y, DUAN H, TIAN X, XU C, WANG W, JIANG W, PANG Z, ZHANG D & TAN Q. 2018. Genetics of obesity traits: A bivariate genome-wide association analysis. Front Genet 9: 179.
• XU B, GOULDING EH, ZANG K, CEPOI D, CONE RD, JONES KR, TECOTT LH & REICHARDT LF. 2003. Brain-derived neurotrophic factor regulates energy balance downstream of melanocortin-4 receptor. Nat Neurosci 6(7): 736-742.
• YEO GSH, FAROOQI IS, AMINIAN S, HALSALL DJ, STANHOPE RG & O’RAHILLY S. 1998. A frameshift mutation in MC4R associated with dominantly inherited human obesity. Nat Genet 20(2): 111-112.
• YU YH. 2017. Making sense of metabolic obesity and hedonic obesity. J Diabetes 9(7): 656-666.
• ZDROJOWY-WEŁNA A, BEDNAREK-TUPIKOWSKA G, ZATOŃSKA K, KOLAČKOV K, JOKIEL-ROKITA A & BOLANOWSKI M. 2020. The association between FTO gene polymorphism rs9939609 and obesity is sex-specific in the population of PURE study in Poland. Adv Clin Exp Med 29(1): 25-32.

Publication Dates

• Publication in this collection
  10 Mar 2023
• Date of issue
  2023

History

• Received
  24 Jan 2022
• Accepted
  2 Aug 2022