README.md

In OchoaLab/simgenphen: Simulate Genotypes and Phenotypes

simgenphen

simgenphen (Simulate Genotypes and Phenotypes) is a wrapper package around several more specialized packages: bnpsd for admixture, simfam for family structure, and simtrait for complex traits. This package assumes reasonable defaults and generally saves on typing for some recurrent cases. Key non-trivial functionality is drawing genotypes for admixed families whose loci are all polymorphic.

Installation

The current development version can be installed from the GitHub repository using devtools:

install.packages("devtools") # if needed
library(devtools)
install_github('OchoaLab/simgenphen')

Examples

There are essentially two recommended ways of using this code, shown in the next two subsections.

library(simgenphen)

One-liner version

Here is a simple way of generating genotypes that feature both admixture and family structure, along with a phenotype! It's run mostly with default values, except data dimensions are smaller:

data <- sim_gen_phen( n_ind = 50, G = 3, m_loci = 100, m_causal = 5 )

# main objects of interest:
# genotype matrix
data$X
# trait vector
data$trait

Version with separate steps

This approach requires more commands, but is more logical and efficient in loops where population structure is constant but want newly-drawn genotype and phenotype replicates. These more explicit commands also return additional values.

First construct population parameters (relatedness but no loci):

k_subpops <- 3
data_pop <- sim_pop( n_ind = 50, G = 3, k_subpops = k_subpops )

Now draw genotypes!

m_loci <- 100
data_geno <- sim_geno(
    data_pop$admix_proportions_1,
    data_pop$inbr_subpops,
    data_pop$fam,
    data_pop$ids,
    m_loci
)

# genotype matrix
data_geno$X

Lastly, draw trait! In this case we specify a non-default heritability and an environment effect "gcat" to showcase that version (those parameters can also be passed to the "one-liner" function sim_gen_phen).

m_causal <- 5
herit <- 0.5
data_trait <- sim_trait_env(
    data_geno$X,
    data_geno$p_anc,
    m_causal,
    herit,
    env = 'gcat',
    k_subpops = k_subpops
)

# the simulated trait
data_trait$trait
# indexes of randomly-selected causal loci
data_trait$causal_indexes
# coefficients of causal loci
data_trait$causal_coeffs

OchoaLab/simgenphen documentation built on Feb. 10, 2023, 3:32 p.m.