Annals of Human Biology (May 2017)
Multilevel modelling of somatotype components: the Portuguese sibling study on growth, fitness, lifestyle and health
Abstract
Background: Somatotype is a complex trait influenced by different genetic and environmental factors as well as by other covariates whose effects are still unclear. Aims: To (1) estimate siblings’ resemblance in their general somatotype; (2) identify sib-pair (brother–brother (BB), sister–sister (SS), brother–sister (BS)) similarities in individual somatotype components; (3) examine the degree to which between and within variances differ among sib-ships; and (4) investigate the effects of physical activity (PA) and family socioeconomic status (SES) on these relationships. Subjects and methods: The sample comprises 1058 Portuguese siblings (538 females) aged 9–20 years. Somatotype was calculated using the Health-Carter method, while PA and SES information was obtained by questionnaire. Multi-level modelling was done in SuperMix software. Results: Older subjects showed the lowest values for endomorphy and mesomorphy, but the highest values for ectomorphy; and more physically active subjects showed the highest values for mesomorphy. In general, the familiality of somatotype was moderate (ρ = 0.35). Same-sex siblings had the strongest resemblance (endomorphy: ρSS > ρBB > ρBS; mesomorphy: ρBB = ρSS > ρBS; ectomorphy: ρBB > ρSS > ρBS). For the ectomorphy and mesomorphy components, BS pairs showed the highest between sib-ship variance, but the lowest within sib-ship variance; while for endomorphy BS showed the lowest between and within sib-ship variances. Conclusions: These results highlight the significant familial effects on somatotype and the complexity of the role of familial resemblance in explaining variance in somatotypes.
Keywords