BayesTwin - An R package for Bayesian Analysis of Twin Data

Statistical analysis of data on twins where inference is focused on determing the relative contribution of nature and nurture to individual differences has been around for a long time. Although the revolution in molecular... [ view full abstract ]

Statistical analysis of data on twins where inference is focused on determing the relative contribution of nature and nurture to individual differences has been around for a long time. Although the revolution in molecular genetics has shifted the focus of the field of behaviour genetics from twin studies to GWAS and GCTA analyses, determing the heritability coefficient for a particular trait remains relevant. Where in the last quarter of the 20^th century, the focus was on using structural equation modelling (SEM) and reporting Maximum Likelihood point estimates and confidence intervals, the first decade of the 21^st century has seen increased use of Bayesian modelling. This has opened up new modelling possibilities such as inference on complex models that were not easily tractable using standard frequentist techniques, including models based on Item Response Theory (IRT) or modelling gene-environment interaction in the presence of gene-environment correlation. However, this new technology with its richness of possibilities has not yet been embraced by the behaviour genetics community, partly due to a lack of standard for reporting results. To make Bayesian analysis more accessible, we introduce the R package BayesTwin. The package includes a wide range of twin models as well as functions that plot relevant information or determine whether the analysis was performed well. In all analyses included in BayesTwin, an IRT model was integrated into the genetic model to facilitate analysis on item level. The integration of such a measurement model is important since an analysis based on an aggregated measure (e.g., a sum-score based analysis) can lead to an underestimation of heritability (van den Berg et al., 2007) or the finding of spurious gene-environment interactions (Schwabe et al., 2014; Molenaar et al., 2014).