Heritability of food preferences in young children

https://doi.org/10.1016/j.physbeh.2006.04.016Get rights and content

Abstract

Background

There is persisting interest in the idea that taste preferences are heritable characteristics, but few twin studies have found evidence for a significant genetic component. Small sample sizes and idiosyncratic selection of foods may have contributed to the negative results. We hypothesized that using a larger twin sample and empirical groupings of food types, would give stronger evidence for the heritability of food preferences.

Objective

We examined the heritability of preferences for four food groups in a sample of young twins.

Design

We administered a food preference questionnaire with 95 foods to 214 mothers of same-sex twin pairs (103 monozygotic and 111 dizygotic pairs) aged 4 to5. 18 foods were excluded because they had been tried by fewer than 25% of the children. Foods were grouped into ‘Vegetables’, ‘Fruits’, ‘Desserts’ and ‘Meat and Fish’ on the basis of a factor analysis of the preference data. Genetic analyses were carried out on mean liking across these four groups, using model fitting techniques.

Results

Over all 77 foods, MZ correlations were higher than DZ correlations for 72 of them, with a higher mean MZ correlation (r = 0.76) than DZ correlation (r = 0.56). Using model fitting techniques with the factor scores, significant heritability estimates were obtained for all four food groups. Heritability was modest for dessert foods (0.20), moderate for vegetables (0.37) and fruits (0.51), and high for liking for protein foods (0.78). Shared environmental effects were strong for desserts, fruits and vegetables, while non-shared environmental influences were low for all four food groups.

Conclusions

These results provide strong evidence for modest heritability of food preferences when using empirically-derived groupings of foods.

Introduction

Taste preferences show remarkable communalities across diverse human populations, with liking for sweetness and dislike for bitterness observed from birth [1], [2], [3]. But nested within the universals are variations in the preferred levels of sweet, sour or bitter tastes as well as liking differences for individual foods [4], [5], [6]. Likes and dislikes are important because they are significant determinants of food acceptance, particularly among children [7], [8], [9].

While universal preferences are assumed to be determined by biological characteristics evolved to promote nutritionally-adequate diets and avoid toxins [3], individual differences could result from environmental or genetic factors. Environmental influences on likes and dislikes are well-established. Nausea can induce aversion for foods eaten shortly before the episode [10]. Observing a disliked food being enjoyed by significant others increases its acceptance [1]. Repeated exposure is also an established method of increasing liking [11], [12], [13].

Genetic influences on taste preferences have been investigated both with ‘bottom-up’ (from biology to behavior) and ‘top-down’ (behavior to biology) methodologies. Bottom-up research has focused on the class of compounds called thioureas which are tasteless to some people but taste bitter to others. A specific genetic locus (taste receptor, type 2, member 38; TAS2R38) on the long arm of chromosome 7 (7q35–36) has been identified for both PTC and PROP [14], [15], which has been associated with the number of fungiform papillae and taste buds on the anterior tongue [16], [17]. PTC and PROP tasters dislike very sweet or bitter foods more than non-tasters [18], [19], [20] and recent results showed PROP tasters to be more sensitive than non-tasters to bitterness and sourness in a range of foods [17]. PTC and PROP tasters have been found to have more food aversions [21] and in some studies they show lower liking for bitter citrus fruits and cruciferous vegetables [22], [23] although not in others [24]. In children, genotypes at the locus related to PROP sensitivity were significantly related to preference for sucrose and for sweet-tasting foods and beverages [5].

The ‘top-down’ approach uses family studies to establish the extent to which familial similarities in food preferences assort with genetic relatedness. A meta-analysis of studies of parent–child pairs concluded that there were significant, but small, parent–child correlations for food preferences [25]. However, because preferences may change over the lifespan, correlations between children and adults could underestimate the familial association. The twin method is a special case of family studies that has been used across many domains of behavior to distinguish environmental and genetic influences. Monozygotic (MZ) twins share all their genes while dizygotic (DZ) twins share on average half their segregating genes. Comparing resemblance between MZ and DZ pairs therefore gives an estimate of the contribution of heritable genetic factors to phenotypic differences. Twin resemblance not attributed to heritable genetic factors is ascribed to shared environmental factors. The remaining reliable phenotypic variance is attributed to non-shared environment, which can be assessed directly by differences within pairs of MZ twins (for details, see [26]).

Contrary to expectation, twin studies of food preferences to date have produced predominantly negative results. Many have found no evidence for a genetic component to food preferences [27], [28], [29], while others [30], [31], [32] have found positive effects for one or two some foods, among many negative results. However, many of the sample sizes were small giving limited power to detect small to moderate levels of heritability. Almost all studies use a single rating of liking for a food as the behavioral measure, which limits the reliability of the estimate of the phenotype. There has also been little or no replication of foods across studies (for example, protein foods have not been adequately assessed), making it difficult to know if small effects have been missed or if heritability varies between foods.

The present study used data from a sample of 214 same-sex twin pairs to investigate the heritability of food preferences in children. Most studies have assessed heritability for individual foods, but we hypothesized that genetic influences may be exerted on preferences for types of foods. We therefore used a larger set of foods than in previous studies and grouped them on the basis of the factor structure of food preferences [33].

Section snippets

Subjects and methods

Participants were drawn from the Twins Early Development Study (TEDS), which includes more than 15,000 pairs of twins born in England and Wales in 1994, 1995 and 1996 [34]. A sub-sample of 214 families with same-sex twins (103 MZ pairs and 111 DZ pairs) aged 4–5 years old took part in the present study. Demographic characteristics of the parents and children are shown in Table 1.

231 families who met the inclusion criteria for the study were invited to take part. 214 families consented and were

Analysis

Intraclass correlations for each of the four food types were compared between MZ and DZ twin pairs. Model fitting techniques using the Mx program [38] were used to apportion the phenotypic variance to genes (A), shared environment (C) or non-shared environment (E). This analysis makes the usual assumptions about the twin method such as the equal environments assumption, genetic additivity, and the absence of assortative mating [26]. The Mx program fits a full ACE model and sub-models to the

Results

The food preference questionnaire had 95 foods, of which 18 had been tried by fewer than 25% of the children and were excluded from the analyses. A comparison of the intraclass correlation coefficients for MZ and DZ twin pairs for each of the remaining food items revealed that MZ correlation coefficients were higher than DZ correlations for 72 out of the 77 foods, with a mean MZ correlation across all foods (r = 0.76) which was substantially higher than the mean DZ correlation (r = 0.56)

Discussion

The results of the present study make a strong case that variations in food preferences are heritable when aggregated across empirically-derived groupings of foods. Using this approach, we found modest heritability estimates for liking for dessert foods, moderate heritability for fruits and vegetables, and high heritability for liking of protein foods (meat and fish). Our positive results contrast with the generally negative findings in the literature, but no other studies have grouped sets of

References (48)

  • G.A. Falciglia et al.

    Evidence for a genetic influence on preference for some foods

    J Am Diet Assoc

    (1994)
  • P. Rozin et al.

    Family environment, not heredity, accounts for family resemblances in food preferences and attitudes: a twin study

    Appetite

    (1987)
  • J. Wardle et al.

    Factor-analytic structure of food preferences in four-year-old children in the UK

    Appetite

    (2001)
  • C.A. Vereecken et al.

    Influence of mother's educational level on food parenting practices and food habits of young children

    Appetite

    (2004)
  • P. Rozin

    The selection of food by rats, humans and other animals

  • L.L. Birch

    Development of food preferences

    Annu Rev Nutr

    (1999)
  • B.J. Cowart

    Development of taste perception in humans: sensitivity and preference throughout the life span

    Psychol Bull

    (1981)
  • P. Rozin et al.

    Food likes and dislikes

    Annu Rev Nutr

    (1986)
  • D.G. Liem et al.

    Heightened sour preferences during childhood

    Chem Senses

    (2003)
  • J.A. Mennella et al.

    Genetic and environmental determinants of bitter perception and sweet preferences

    Pediatrics

    (2005)
  • H.A. Raynor et al.

    Is dietary fat intake related to liking or household availability of high- and low-fat foods?

    Obes Res

    (2004)
  • E. Bere et al.

    Correlates of fruit and vegetable intake among Norwegian schoolchildren: parental and self-reports

    Public Health Nutr

    (2004)
  • J.O. Fisher et al.

    Parental influences on young girls' fruit and vegetable, micronutrient, and fat intakes

    J Am Diet Assoc

    (2002)
  • S.A. Sullivan et al.

    Infant dietary experience and acceptance of solid foods

    Pediatrics

    (1994)
  • Cited by (0)

    Funding: The research was funded by a grant from the Biotechnology and Biological Sciences Research Council (31/D19086). Jane Wardle is supported by Cancer Research UK. The TEDS project is funded by the Medical Research Council (G0500079).

    View full text