gnomesims.RmdThe gnomesims package is designed to provide estimates
of gene-environment correlation for simulated data. It is a tool for
researchers who use polygenic scores of twins, parents and siblings to
detect gene-environment correlation and want to address issues of power,
sample size and effect size. We focus on two types of gene-environment
correlation, namely, cultural transmission (= genetic nurture) and
sibling interaction.
The package can be installed from its’ Github repository using
devtools.
# install.packages("devtools")
library(devtools)
devtools::install_github("josefinabernardo/gnomesims")Next, it should be loaded it into you current session.
library(gnomesims)The core function of gnomesims is the
gnome_mx_simulation() function. It takes in the ACE
estimates, sample sizes, and effect size measures as arguments and
returns two data frames with power estimates and path coefficients.
gnome_mx_simulation(ct = .01, si = .025, npgsloci = 10)
## [1] "Running simulation proportion of genetic variance explained by the PGS is: 0.1 ."
## [1] "The factorial design has 1 setting(s)."
## [1] 1
## $power
## nmz ndz a c e g b x PGS A p1 p2 p3
## 1 4000 4000 0.6324555 0.5477226 0.5477226 0.01 0.025 0 0.1 0.9 0.13 0.181 0.06
## p4 p5 p6 p7 p8 Smz Sdz
## 1 0.108 0.122 0.149 0.054 0.079 0.04559689 0.0297855
##
## $params
## nmz ndz a c e g b x PGS A e1 e2 e3
## 1 4000 4000 0.6324555 0.5477226 0.5477226 0.01 0.025 0 0.1 0.9 0.023 0.031 0.01
## e4 e5 e6 e7 e8 Smz Sdz
## 1 0.025 0.029 0.032 0.01 0.025 0.04559689 0.0297855Embedded within the simulation is the function
gnome_power() to calculate power from alpha, degrees of
freedom and the non-centrality parameter.
We recognize that the path coefficients are difficult to interpret.
To solve this issue, the function gnome_effect() calculates
a readily interpretable effect size measure.
It is possible to look at results for present but unmodelled assortative mating by specifying the genotypic correlation between the parents. We are working on a version that estimates direct assortative mating when fitting the model.
gnome_mx_simulation(ct = .01, si = .025, npgsloci = 10, assortm = .26)
## [1] "Running simulation proportion of genetic variance explained by the PGS is: 0.1 ."
## [1] "The factorial design has 1 setting(s)."
## [1] 1
## $power
## nmz ndz a c e g b x PGS A assortm p1
## 1 4000 4000 0.6324555 0.5477226 0.5477226 0.01 0.025 0 0.1 0.9 0.26 0.141
## p2 p3 p4 p5 p6 p7 p8 Smz Sdz
## 1 0.189 0.06 0.106 0.132 0.157 0.054 0.078 0.04559689 0.0297855
##
## $params
## nmz ndz a c e g b x PGS A assortm e1
## 1 4000 4000 0.6324555 0.5477226 0.5477226 0.01 0.025 0 0.1 0.9 0.26 0.024
## e2 e3 e4 e5 e6 e7 e8 Smz Sdz
## 1 0.031 0.01 0.025 0.029 0.032 0.01 0.025 0.04559689 0.0297855The package can also simulate results using generalized estimating
equations (gee) with the gnome_gee_simulation() function.
Functionality and results are similar to those of the
gnome_mx_simulation() function.
gnome_gee_simulation(ct = .01, si = .025, npgsloci = 10)
## [1] "Running simulation proportion of genetic variance explained by the PGS is: 0.1 ."
## [1] "The factorial design has 1 setting(s)."
## [1] 1
## $power
## nmz ndz a c e g b x PGS A p1 p2 p3
## 1 4000 4000 0.6324555 0.5477226 0.5477226 0.01 0.025 0 0.1 0.9 0.137 0.173 0.06
## p4 p5 p6 p7 p8 Smz Sdz
## 1 0.105 0.12 0.143 0.054 0.077 0.04559689 0.0297855
##
## $params
## nmz ndz a c e g b x PGS A e1 e2 e3
## 1 4000 4000 0.6324555 0.5477226 0.5477226 0.01 0.025 0 0.1 0.9 0.025 0.061 0.01
## e4 e5 e6 e7 e8 Smz Sdz
## 1 0.05 0.029 0.063 0.01 0.05 0.04559689 0.0297855To demonstrate what type of data this package can generate, we have
included two in-built data sets. They contain the results of the
gnome_mx_simulation() function for 3 x 3 = 9 combination of
AC covariance input parameters. The data set
gnome_power_data contains the power results and the data
set gnome_params_data contains parameter estimates.
gnome_power_data
## nmz ndz a c e CT SI x PGS A CT(m1) MZDZ SI(m2) MZDZ
## 1 4000 4000 0.63 0.55 0.55 0.00 0.00 0 0.1 0.9 0.050 0.050
## 2 4000 4000 0.63 0.55 0.55 0.05 0.00 0 0.1 0.9 0.408 0.161
## 3 4000 4000 0.63 0.55 0.55 0.10 0.00 0 0.1 0.9 0.917 0.474
## 4 4000 4000 0.63 0.55 0.55 0.00 0.05 0 0.1 0.9 0.160 0.399
## 5 4000 4000 0.63 0.55 0.55 0.05 0.05 0 0.1 0.9 0.756 0.754
## 6 4000 4000 0.63 0.55 0.55 0.10 0.05 0 0.1 0.9 0.989 0.945
## 7 4000 4000 0.63 0.55 0.55 0.00 0.10 0 0.1 0.9 0.505 0.929
## 8 4000 4000 0.63 0.55 0.55 0.05 0.10 0 0.1 0.9 0.953 0.992
## 9 4000 4000 0.63 0.55 0.55 0.10 0.10 0 0.1 0.9 0.999 0.999
## CT(m3) MZDZ SI(m3) MZDZ CT(m1) DZ SI(m2) DZ CT(m3) DZ SI(m3) DZ Smz Sdz
## 1 0.050 0.050 0.050 0.050 0.050 0.050 0.00% 0.00%
## 2 0.300 0.050 0.261 0.156 0.152 0.050 6.82% 6.82%
## 3 0.787 0.050 0.724 0.453 0.424 0.050 14.65% 14.65%
## 4 0.050 0.292 0.179 0.302 0.050 0.171 6.57% 3.41%
## 5 0.285 0.281 0.640 0.651 0.146 0.163 13.90% 10.74%
## 6 0.761 0.271 0.940 0.894 0.404 0.156 22.22% 19.06%
## 7 0.050 0.800 0.550 0.819 0.050 0.506 13.65% 7.32%
## 8 0.270 0.783 0.915 0.962 0.140 0.480 21.47% 15.15%
## 9 0.734 0.763 0.994 0.995 0.384 0.454 30.30% 23.97%