Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Likelihood ratio tests in behavioral genetics: problems and solutions.

Annica Dominicus1, Anders Skrondal, Håkon K Gjessing

  • 1Department of Mathematics, Stockholm University, Stockholm, Sweden. annicad@math.su.se

Behavior Genetics
|February 14, 2006
PubMed
Summary

The likelihood ratio test for family data requires adjusted p-values when testing variance components. Using a mixture of chi-square distributions provides accurate results for genetic and environmental influence models.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Comorbidities in parents of individuals with congenital heart block: a Swedish nationwide study.

Annals of the rheumatic diseases·2026
Same author

Longitudinal study of genome-wide DNA methylation in individuals with and without post-acute symptoms following SARS-CoV-2 infection.

Epigenomics·2026
Same author

Nationwide Temporal Trends in Adverse Pregnancy Outcomes and Treatments in Systemic Lupus Erythematosus Pregnancy Over Two Decades in Sweden.

Arthritis & rheumatology (Hoboken, N.J.)·2025
Same author

Statistical methods to disentangle genetic effects influencing infertility and early fetal viability with a genome-wide application.

PLoS genetics·2025
Same author

Maternal and Fetal SERPINA3 Polymorphisms and Risk of Preeclampsia: A Dyad and Triad Based Case-Control Study.

Current issues in molecular biology·2025
Same author

Thiopurines and Intrahepatic Cholestasis of Pregnancy in Systemic Lupus Erythematosus and Inflammatory Bowel Disease Pregnancies.

The American journal of gastroenterology·2025

Area of Science:

  • Quantitative genetics
  • Statistical genetics
  • Behavioral genetics

Background:

  • Likelihood ratio tests (LRTs) are crucial for assessing genetic and environmental influences on traits using family data.
  • Standard LRTs assume a chi-square distribution, which is inaccurate when variance components are non-negative, leading to inflated p-values and problematic model selection (AIC, BIC).

Purpose of the Study:

  • To adapt existing theory for LRTs of nested models in family data, specifically addressing boundary issues with non-negative variance components.
  • To derive accurate asymptotic distributions and mixing probabilities for LRT statistics in classical twin models.

Main Methods:

  • Focusing on the ACE twin model for univariate traits, the study adapted theory to determine the asymptotic distribution of the LRT statistic.

Related Experiment Videos

  • Derived mixing probabilities for chi-square distributions to correct p-values when testing nested models.

    • Main Results:

    • The asymptotic distribution of the LRT statistic is a mixture of chi-square distributions.
    • For testing AE or CE against ACE models, p-values using chi-square with 1 df should be halved.
    • Testing the E model against ACE requires a mixture of chi-square distributions with 0, 1, and 2 df, with derived mixing probabilities.

    Conclusions:

    • Standard LRT procedures yield inaccurate p-values for models with non-negative variance components.
    • Failure to use the derived mixture of chi-square distributions as the reference distribution can lead to invalid conclusions regarding genetic and environmental influences.
    • The study provides specific adjustments for testing nested models within the ACE and ADE frameworks.