American Journal of Human Biology · 2018 · 16 citations · 74 references
Life SciencesNutritionNutritional SciencesGenetic EpidemiologyHuman GeneticsPopulation HealthFamily-based Genetic AnalysisGene-environment InteractionHeritabilitySocioeconomic Context Of FamiliesSocioeconomic Determinants Of HealthSocioeconomic Determinants Of ObesityBayesian Linear Mixed ModelsFamily Socioeconomic StatusFamily-based HeritabilityHeritability Of Anthropometric TraitsHeritability Of Body Weight TraitsLinear Mixed ModelsNuclear Family StructureNutritional Trait Heritability
OBJECTIVES: General health status is reflected in measures of height, weight, and BMI. Assessing sources of variation in these outcomes reveals population-specific variables of importance to health and nutrition. We characterize the impacts of socioeconomic variables related to the nuclear family on health outcomes of boat-dwelling Shodagor children, mothers, and fathers, and to estimate the proportion of variation in height, weight, and BMI influenced by both genetic variation and nongenetic variation among household environments. METHODS: Bayesian linear mixed models (LMMs) estimate heritability and household-effect variance components among the Shodagor. These models also assess the influences of specific socioeconomic predictor variables on different types of individuals within the household (children, mothers, and fathers). RESULTS: Overall, models explain 61.7% of variation in height, 59.4% in weight, and 65.8% in BMI for this sample of Shodagor. Mother's decision-making and household income have expected, positive associations with children's weight and BMI. Number of children has an unexpected positive relationship to children's height and a negative relationship to father's BMI. Genetic variation explains less than 26% of phenotypic variation for each of these traits on average. CONCLUSIONS: Our results show that resource flows and distributions within Shodagor households account for a significant amount of variance in nutritional outcomes. Problems commonly associated with increasing market integration may lead to negative outcomes for children, while mother's autonomy may lead to positive outcomes. Our models also indicate that environmental factors account for more variation in these outcomes than expected, relative to genetics, and we discuss the implications.
74
R: A Language and Environment for Statistical Computing
R Core Team · 2000 · 353K citations · Full text
Natural Sciences, Computer Science, Statistical Computing +10