| 1. |
- Dominicus, Annica, et al.
(författare)
-
Bias in variance components due to nonresponse in twin studies
- 2006
-
Ingår i: Twin Research and Human Genetics. - 1832-4274 .- 1839-2628. ; 9:2, s. 185-193
-
Tidskriftsartikel (refereegranskat)abstract
- Incomplete data on trait values may bias estimates of genetic and environmental variance components obtained from twin analyses. If the nonresponse mechanism is 'ignorable' then methods such as full information maximum likelihood estimation will produce consistent variance component estimates. If, however, nonresponse is 'nonignorable', then the situation is more complicated. We demonstrate that a within-pair correlation of nonresponse, possibly different for monozygotic (MZ) and dizygotic (DZ) twins, may well be compatible with 'ignorability'. By means of Monte Carlo simulation, we assess the potential bias in variance component estimates for different types of nonresponse mechanisms. The simulation results guide the interpretation of analyses of data on perceptual speed from the Swedish Adoption/Twin Study of Aging. The results suggest that the dramatic decrease in genetic influences on perceptual speed observed after 13 years of follow-up is not attributable solely to dropout from the study, and thus support the hypothesis that genetic influences on some cognitive abilities decrease with age in late life.
|
|
| 2. |
- Dominicus, Annica, 1976-
(författare)
-
Latent variable models for longitudinal twin data
- 2006
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Longitudinal twin data provide important information for exploring sources of variation in human traits. In statistical models for twin data, unobserved genetic and environmental factors influencing the trait are represented by latent variables. In this way, trait variation can be decomposed into genetic and environmental components. With repeated measurements on twins, latent variables can be used to describe individual trajectories, and the genetic and environmental variance components are assessed as functions of age. This thesis contributes to statistical methodology for analysing longitudinal twin data by (i) exploring the use of random change point models for modelling variance as a function of age, (ii) assessing how nonresponse in twin studies may affect estimates of genetic and environmental influences, and (iii) providing a method for hypothesis testing of genetic and environmental variance components. The random change point model, in contrast to linear and quadratic random effects models, is shown to be very flexible in capturing variability as a function of age. Approximate maximum likelihood inference through first-order linearization of the random change point model is contrasted with Bayesian inference based on Markov chain Monte Carlo simulation. In a set of simulations based on a twin model for informative nonresponse, it is demonstrated how the effect of nonresponse on estimates of genetic and environmental variance components depends on the underlying nonresponse mechanism. This thesis also reveals that the standard procedure for testing variance components is inadequate, since the null hypothesis places the variance components on the boundary of the parameter space. The asymptotic distribution of the likelihood ratio statistic for testing variance components in classical twin models is derived, resulting in a mixture of chi-square distributions. Statistical methodology is illustrated with applications to empirical data on cognitive function from a longitudinal twin study of aging.
|
|
| 3. |
- Dominicus, Annica, et al.
(författare)
-
Likelihood ratio tests in behavioral genetics: Problems and solutions
- 2006
-
Ingår i: Behavior Genetics. - : Springer. - 0001-8244 .- 1573-3297. ; 36:2, s. 331-340
-
Tidskriftsartikel (refereegranskat)abstract
- The likelihood ratio test of nested models for family data plays an important role in the assessment of genetic and environmental influences on the variation in traits. The test is routinely based on the assumption that the test statistic follows a chi-square distribution under the null, with the number of restricted parameters as degrees of freedom. However, tests of variance components constrained to be non-negative correspond to tests of parameters on the boundary of the parameter space. In this situation the standard test procedure provides too large p-values and the use of the Akaike Information Criterion (AIC) or the Bayesian Information Criterion (BIC) for model selection is problematic. Focusing on the classical ACE twin model for univariate traits, we adapt existing theory to show that the asymptotic distribution for the likelihood ratio statistic is a mixture of chi-square distributions, and we derive the mixing probabilities. We conclude that when testing the AE or the CE model against the ACE model, the p-values obtained from using the v2 (1 df) as the reference distribution should be halved. When the E model is tested against the ACE model, a mixture of v2(0 df), v2(1 df) and v2 (2 df) should be used as the reference distribution, and we provide a simple formula to compute the mixing probabilities. Similar results for tests of the AE, DE and E models against the ADE model are also derived. Failing to use the appropriate reference distribution can lead to invalid conclusions.
|
|
| 4. |
- Holmberg, Lars, et al.
(författare)
-
Prognostic markers under watchful waiting and radical prostatectomy
- 2006
-
Ingår i: Hematology/Oncology Clinics of North America. - : Elsevier. - 0889-8588 .- 1558-1977. ; 20:4, s. 845-855
-
Tidskriftsartikel (refereegranskat)abstract
- A suitable setting to analyze factors that determine prognosis or treatment response in prostate cancer is an unbiased comparison of radical prostatectomy and watchful waiting as in the Scandinavian Prostate Cancer Group Trial number 4. In our previous presentation of 10-year results, we studied Gleason score, serum prostate-specific antigen (PSA) at diagnosis, and age at diagnosis as modifiers of the effect of radical prostatectomy on survival. Because overall prognostic information obtained by these parameters or by tumor stage was not provided in our publication, we now present these data in the two study arms separately.
|
|
| 5. |
- Smedby, Karin Ekström, et al.
(författare)
-
Variation in DNA repair genes ERCC2, XRCC1, and XRCC3 and risk of follicular lymphoma
- 2006
-
Ingår i: Cancer Epidemiology, Biomarkers and Prevention. - 1055-9965 .- 1538-7755. ; 15:2, s. 258-265
-
Tidskriftsartikel (refereegranskat)abstract
- The reasons for the positive association between skin cancer and non-Hodgkin's lymphoma are not known but may be due to common susceptibility involving suboptimal DNA repair. Therefore, we investigated selected polymorphisms and haplotypes in three DNA repair genes, previously associated with skin cancer and DNA repair capacity, in risk of follicular lymphoma, including possible gene interaction with cigarette smoking and sun exposure. We genotyped 19 single nucleotide polymorphisms (SNP) in the ERCC2, XRCC1, and XRCC3 genes in 430 follicular lymphoma patients and 605 controls within a population-based case-control study in Denmark and Sweden. Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated using unconditional logistic regression and haplotype associations were assessed with a global score test. We observed no associations between variation in the ERCC2 and XRCC1 genes and follicular lymphoma risk. In XRCC3, increased risk of follicular lymphoma was suggested for rare homozygotes of three SNPs [Rs3212024: OR, 1.8 (95% CI, 1.1-2.8); Rs3212038: OR, 1.5 (95% CI, 1.0-2.4); Rs3212090: OR, 1.5 (95% CI, 1.0-2.5)]. These results were strengthened in current cigarette smokers. However, evidence of differences in XRCC3 haplotype distributions between follicular lymphoma cases and controls was weak, both overall and in current smokers. We conclude that polymorphic variation in the XRCC3 gene, but not in ERCC2 or XRCC1, may be of importance for susceptibility to follicular lymphoma, perhaps primarily in current smokers. The link between skin cancer and follicular lymphoma is unlikely to be mediated through common variation in the studied DNA repair gene polymorphisms.
|
|
