In qPharmetra/qpToolkit: qPharmetra R Tools

library(qpToolkit)
library(ggplot2)
library(Hmisc)

Kick off by creating some play data with PK and PD data.

pkpdData = example.pkpdData()

example panel functions

example("panel.thalf")
example("panel.modelfit")
example("panel.cwres")
example("panel.residual")
example("panel.superpose.arrows")
example("panel.superpose.bubble")

example("plot.tornado")

example VPC plot using xpose4

example.xpose.VPC()
library(xpose4)
xpose.VPC(
   system.file(package = 'qpToolkit','NONMEM/vpc_final_strt/vpc_results.csv'),
   system.file(package = 'qpToolkit','NONMEM/vpc_final_strt/vpctab'), 
   logy=T,, by='STRT',
       col=grey(0.4),  cex = 1,
       PI.ci.area.smooth = T,
       PI.ci.med.arcol = PI.ci.med.arcol,
       PI.real.med.col = PI.real.med.col,
       PI.real.down.col = PI.real.down.col,
       PI.real.up.col = PI.real.up.col,
       PI.ci.down.arcol = PI.ci.down.arcol,
       PI.ci.up.arcol = PI.ci.up.arcol
   )

example VPC plot using qPharmetra's ggvpc

vpc = nm.read.vpc(system.file(package = 'qpToolkit','/NONMEM/vpc_final_strt'), PI.ci.area.smooth=T)

ggvpc_xpose(vpc, point.size = 2) +
   facet_grid(~strata) +
   scale_y_log10()

ggvpc_standard(vpc, point.size = 2) +
   facet_grid(~strata) +
    scale_y_log10()

ggvpc functions can plot by multiple gropuping levels in 'facet_grid' mode

Since we do not have data available, we quickly scramble some data, duplicate and tweak it

vpc = nm.read.vpc(system.file(package = 'qpToolkit','/NONMEM/vpc_final_strt'), PI.ci.area.smooth=T)

vpc = lapply(vpc, function(x) expand.data(x, values=1:2,name = "pat.class"))

vpc$obs$TIME[vpc$obs$pat.class==1] = vpc$obs$TIME[vpc$obs$pat.class==1]/0.5
vpc$res$xCov[vpc$res$pat.class==1] = vpc$res$xCov[vpc$res$pat.class==1]/0.5
vpc$vpc$xCovm[vpc$vpc$pat.class==1] = vpc$vpc$xCovm[vpc$vpc$pat.class==1]/0.5

vpc$obs$TIME[vpc$obs$pat.class==2] = vpc$obs$TIME[vpc$obs$pat.class==2]/1.5
vpc$res$xCov[vpc$res$pat.class==2] = vpc$res$xCov[vpc$res$pat.class==2]/1.5
vpc$vpc$xCovm[vpc$vpc$pat.class==2] = vpc$vpc$xCovm[vpc$vpc$pat.class==2]/1.5

vpc = lapply(vpc, function(x) {
   x$pat.class = swap(x$pat.class, sunique(x$pat.class), Cs(Obese,Normal_BodyWeight))
   return(x)}
   )

vpc = lapply(vpc, function(x) {
   x$strata = swap(x$strata, sunique(x$strata), Cs(study1,study2))
   return(x)}
   )

ggvpc_xpose(vpc, point.size = 2) +
   facet_grid(pat.class~strata) +
   scale_y_log10()