Refit fastlmm model with new delta

Refit fastlmm model with new delta value

Usage

refitModel(fit, delta = NULL, interceptOnly = FALSE, fixedNBtheta = FALSE)

Arguments

fit

model fit of class fastlmm or fastglmm

delta

new value for delta

interceptOnly

if TRUE, only retain the intercept from the fixed effect design matrix

fixedNBtheta

if FALSE, low theta in NB model to be re-estimated too

Details

Useful for evaluating log-likelihood and multiple values of delta

Examples

library(lme4)

fit <- fastlmm(Reaction ~ Days + (1 | Subject), sleepstudy)

summary(fit)
#> Linear mixed model fit by ML ['fastlmm']
#>  Formula: Reaction ~ Days + (1 | Subject)
#> 
#> Coefficients:
#>             Estimate Std. Error    df t value Pr(>|t|)    
#> (Intercept) 251.4051     9.5062  24.5   26.45   <2e-16 ***
#> Days         10.4673     0.8017 162.0   13.06   <2e-16 ***
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> Residual df: 162.2 
#> 
#> Variance components:
#>   sigSq_g: 1297
#>   sigSq_e: 954.5
#>   delta:   0.736

summary(refitModel(fit, delta=1000))
#> Linear mixed model fit by ML ['fastlmm']
#>  Formula: Reaction ~ Days + (1 | Subject)
#> 
#> Coefficients:
#>             Estimate Std. Error    df t value Pr(>|t|)    
#> (Intercept)  251.405      6.563 127.9  38.308  < 2e-16 ***
#> Days          10.467      1.228 162.0   8.527 1.01e-14 ***
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> Residual df: 177.8 
#> 
#> Variance components:
#>   sigSq_g: 2.238
#>   sigSq_e: 2238
#>   delta:   1000