Constraint on variance components - reworkhow/JWAS.jl GitHub Wiki

In multi-trait analysis, the constraint can be applied to variance components to make the covariance zero. In this situation, the variance will be sampled as if from a single trait for each trait separately. An error message will appear if any constraint is used in the single-trait analysis.

Constraint on residual variance

Add argument constraint=true in build_model() function. By default, constraint=false.

Example:

model = build_model(model_equation,constraint=true)

Please also find this printed information in JWAS: Constraint on residual variance-covariance matrix (i.e., zero covariance).

Constraint on marker effect/genetic variance:

Add argument constraint=true in get_genotypes() function. By default, constraint=false.

Example:

genotypes  = get_genotypes(genofile,separator=',',method="BayesC",constraint=true)

Please also find this printed information in JWAS: Constraint on marker effect variance-covariance matrix (i.e., zero covariance) for genotypes.