Twin Data

Francisco Cardozo

University of Miami

2022-12-06

Twin data

  • Twin data are dyadic data (Non-independence)
  • Twins are more similar than a random person in the data set.
  • Twins models can separate variance into genetic and environmental sources.

Dyadic Data Analsyis: SEM approach

Variance Components

  • Poposed orignailly by Shaun Purcell

    • Additive variation in genotype (A)
    • Shared environment (C)
    • Nonshared environment (E)

  • Moderation:

“The parameters of interest (e.g., a2, c2, and e2) are estimated for”exposed” and “unexposed” individuals separately.”

“A test of G × E is given by equating the parameters across exposure groups and observing the associated decline in model fit.”

Example

“Our investigation supports our hypothesis that the magnitude of genetic influences on cognitive aptitude varies with socioeconomic status. … Our results suggest that differences between middle-class and affluent parents affect the expression of genetic potential.”

The models suggest that in impoverished families, 60% of the variance in IQ is accounted for by the shared environment, and the contribution of genes is close to zero; in affluent families, the result is almost exactly the reverse.

Turkheimer, et al. 

Harden, et al. 

  • They wanted to study the genes-environment interaction
  • They changed the SES measure to parents’ income and education
  • They have two different types of twin
    • MZ monozygotic.
    • DZ dizygotic.
  • They estimated the following model.

Genes Enviroment Interaction (GxE)

The model

ACE are latent variables with mean = 0, SD = 1, and the squared path coefficient from the A component to the phenotype == the heritability. The GxE model estimates that path coefficient as the function of the measured environmental variable (ENV): a + a’ X env, such that a2 is the heritability when ENV ==0 (if it has been standardized to M = 0, SD = 1, then at the mean of the environmental variable) and a’ is the interaction parameter.

Results

Results 2

Other models

More models

Factor Mixture Analysis

More models

Factor Mixture ACE analysis

More models

  • ACDE
  • AC’RE
    • The common environments (C) are partitioned into a rearing environment (R) and an extended community or family environment (C’).

Potential application

  • Analysis of dyadic data.
Parent Student n
1 Female Female 114
2 Female Male 55
3 Male Female 16
4 Male Male 9

Potential application

Sex differences in perception of risk?

Thanks

EPH 701 Social Genomics