Hostname: page-component-7c8c6479df-7qhmt Total loading time: 0 Render date: 2024-03-27T11:47:06.085Z Has data issue: false hasContentIssue false

The Heritability of Breast Cancer: A Bayesian Correlated Frailty Model Applied to Swedish Twins Data

Published online by Cambridge University Press:  21 February 2012

Isabella Locatelli*
Affiliation:
University Luigi Bocconi, Milan, Italy. isabella.locatelli@uni-bocconi.it
Paul Lichtenstein
Affiliation:
Karolinska Institut, Stockholm, Sweden.
Anatoli I. Yashin
Affiliation:
Max Planck Institute for Demographic Research, Rostock, Germany.
*
*Address for correspondence: Isabella Locatelli, Via Mincio 30, 20139 MILANO, Italy.*

Abstract

Core share and HTML view are not available for this content. However, as you have access to this content, a full PDF is available via the ‘Save PDF’ action button.

The aim of this study was to investigate the role of genes and environment in susceptibility to breast cancer and to give an estimate of heritability in the propensity to develop the disease. To do this we applied an interdisciplinary approach, merging models developed in the field of demography and survival analysis — so-called frailty models — and models coming from quantitative genetics and epidemiology, namely genetic models. In our study, the inferential problem was solved in a Bayesian framework and the numerical work was carried out using MCMC methods. We used the special information coming from twin data, particularly breast cancer data, from the Swedish Twin Register. The application of a correlated log-normal frailty model leads to a very large estimate of the population heterogeneity (σ = 6.7), and relatively small correlations between co-twins' frailties — around 0.3 for monozygotic and 0.1 for dizygotic twins. Comparing three different genetic models (an ACE, an AE and an ADE model), we furthermore concluded that genetic effects would explain globally almost 30% of the total variability of propensity to breast cancer. Environmental effects would be predominant in determining breast cancer susceptibility and these effects would be primarily individual-specific, that is, non-shared effects. Finally, a model which includes dominance genetic effects (ADE model) is preferred for genetic and statistical reasons.

Type
Articles
Copyright
Copyright © Cambridge University Press 2004