rxode2lincmt
Linear Compartment Model Solutions and Gradients for 'rxode2'
Analytic one, two and three compartment linear pharmacokinetic
solutions with their parameter gradients from 'stan' automatic
differentiation (Carpenter et al (2015)
rxode2lincmt
rxode2lincmt provides the analytic one, two and three compartment
linear pharmacokinetic solutions and their parameter gradients that
rxode2's linCmt() uses. The
gradients come from 'stan' automatic differentiation.
This code used to be compiled inside 'rxode2'. It lives in its own
package so that installing 'rxode2' does not compile 'stan': 'rxode2'
imports rxode2lincmt and calls the compiled kernels through an
external-pointer table. Most users never call rxode2lincmt directly;
they write linCmt() in an 'rxode2' or 'nlmixr2' model.
The closed-form solutions follow the idea of the wnl package by Kyun-Seop Bae, though the implementation here is different.
Installation
You can install the development version of rxode2lincmt from
GitHub with:
# install.packages("devtools")
devtools::install_github("nlmixr2/rxode2lincmt")
Compiling the 'stan'-based solutions takes a while.
Examples
The eigenvalues and coefficient matrices of a two compartment model from its micro-constants:
library(rxode2lincmt)
solComp2(k10 = 0.1, k12 = 3, k21 = 1)
#> $L
#> [1] 4.07546291 0.02453709
#>
#> $C1
#> [,1] [,2]
#> [1,] 0.7592000 0.2408000
#> [2,] -0.7405714 0.7405714
#>
#> $C2
#> [,1] [,2]
#> [1,] -0.2468571 0.2468571
#> [2,] 0.2408000 0.7592000
The micro-constants of a two compartment model given clearance and volume parameters:
linCmtMicros(p1 = 2, v1 = 20, p2 = 3, p3 = 40, ncmt = 2)
#> v k k12 k21
#> 20.000 0.100 0.150 0.075
One step of the per-row kernel: the concentration one time unit after a
100-unit bolus into a one compartment model with clearance 2 and volume
20, which matches the closed form 100 / 20 * exp(-2 / 20):
step <- linCmtModelDouble(dt = 1, p1 = 2, v1 = 20, p2 = 0, p3 = 0, p4 = 0, p5 = 0,
ka = 0, alastNV = 100, rateNV = 0, ncmt = 1L, oral0 = 0L,
trans = 1L, deriv = FALSE, type = 0L, tau = 0, tinf = 0,
amt = 0, bolusCmt = 0L, ndiff = 0L)
step$val
#> [1] 4.524187
100 / 20 * exp(-2 / 20)
#> [1] 4.524187