computeVPC.NB {HeritSeq}

Calculate the negative binomial (NB) variance partition coefficient (VPC) for one or more features.

Description

Calculate the NB VPC for one or more features following the model fitting function fit.NB().

Usage

computeVPC.NB(para)

Arguments

para

A G \times 3 matrix of negative binomial fit parameters for G features, G\geq 1. The column order is intercept \alpha_g, random effect \sigma_g^2 (\sigma_g^2\geq0), and dispersion \phi (\phi>0).

Value

A G \times 1 matrix consisting of VPC for G features based on negative binomial mixed model. Column name is "NB-fit"; row names are the feature names.

Examples

## Compute VPC for each feature under negative binomial mixed model.
vpc.nb <- computeVPC.NB(para_nb)

## Visulize the distribution of the VPCs. 
hist(vpc.nb, breaks = 50, col = "cyan")

## Plot sorted VPCs.
plot(sort(vpc.nb), ylab = "Heritability (h2)", ylim = c(0,1), 
main = "Sorted NB VPC scores")
abline(h = 0.9, lty = 2, col = "red")
text(50, 0.92, "h2 = 0.9", col = "red")