Giters

andrewhooker / PopED

Population Experimental Design (PopED) in R

Home Page:https://andrewhooker.github.io/PopED/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

unused argument in plot_model_prediction()

UNCcrawler opened this issue · comments

commented

Hi Professor Hooker:
I am new to your package. I was trying to prepare a template using RxODE and encountered two unexpected errors. Could you please help me?

  1. The error is in bold
    plot_model_prediction(design$poped.db, model_num_points = 500, PI = T)
    Error in model_prediction(poped.db, model_num_points = model_num_points, :
    unused argument (PI = T)
    In addition, can we change the display of the plot? eg. color, label, legend title, etc

  2. In the output, the parameters were not named as defined in create.poped.database(d = c(KA = 0.09, CL = 0.25^2, V = 0.09)) as in your example
    optimized.design$rse
    bpop[1] bpop[2] bpop[3] D[1,1] D[2,2] D[3,3]
    13.181201 4.887268 9.723204 95.296309 33.934192 51.834468
    SIGMA[1,1] SIGMA[2,2]
    39.510102 26.998621

The full code is below:
`library(PopED)
library(RxODE)

sessionInfo()

########################################################################################################

Define model using RxODE

########################################################################################################

model structure

mod = RxODE({
CP = CENT/V;
d/dt(DEPOT) = -KADEPOT;
d/dt(CENT) = KADEPOT - (CL/V)*CENT;
})

simulation function

xt=sampling time, p=parameter, poped.db = PopED database

ff = function(model_switch, xt, p, poped.db){
times_xt = drop(xt)
ev = et(time = 0, amt = p[["DOSE"]], ii = p[["TAU"]], until = max(times_xt))%>%
et(times_xt)

out = rxSolve(mod, p, ev)
return(list(y = matrix(out$CP, ncol =1), poped.db = poped.db))
}

define BSV

a = c(dose, dosing interval), bpop = theta, b = eta

sfg = function(x, a, bpop, b, bocc){
parameters = c(
KA = bpop[1]*exp(b[1]),
CL = bpop[2]*exp(b[2]),
V = bpop[3]*exp(b[3]),
DOSE = a[1],
TAU = a[2]
)
return(parameters)
}

define residual error

feps = function(model_switch, xt, parameters, epsi, poped.db){
y = do.call(poped.db$model$ff_pointer, list(model_switch, xt, parameters, poped.db))[[1]] # simulate to obtain F
y = y*(1+epsi[,1])+epsi[,2]
return(list(y = y, poped.db = poped.db))
}

create PopED databases

poped_db = create.poped.database(ff_fun = ff,
fg_fun = sfg,
fError_fun = feps,
bpop = c(KA = 0.25, CL = 3.75, V = 72.8), # population parameter
d = c(KA = 0.09, CL = 0.25^2, V = 0.09), # BSV variability in variance
sigma = c(prop = 0.04, add = 0.0025),
m = 2, # 2 groups
groupsize = 20, # each group has 20 subjects
xt = c(1, 2, 8, 240, 245), # initial design of sampling time
minxt = c(0, 0, 0, 240, 240), # 5 sampling timepoints, minimum for each
maxxt = c(10, 10, 10, 248, 248), # 5 sampling timespoints, maximum for each
bUseGrouped_xt = 1, # set both groups have same sampling time to true
a = cbind(DOSE = c(20, 40), TAU = c(24, 24)), # initial dosing for group 1 (dose 20, tau 24) and group 2 (dose 40, tau 24)
maxa = c(DOSE = 200, TAU = 24), # maximum dose and dosing interval
mina = c(DOSE = 0, TAU = 24)) # minimum dose and dosing interval

model prediction

pred = model_prediction(poped_db,
model_num_points = 500, # result in 505 observation for each group (+5 xt time points)
include_sample_times = T)

plot model prediction

plot_model_prediction(poped_db,
model_num_points = 500,
separate.groups = T)

plot_model_prediction(poped_db,
model_num_points = 500,
separate.groups = T,
PI = T)

evaluate design

intial.design = evaluate_design(poped_db)
intial.design$ofv
intial.design$rse

design optimization, time as continuous

design = poped_optim(poped_db, opt_xt = T)
plot_model_prediction(design$poped.db, model_num_points = 500)
optimized.design = evaluate_design(design$poped.db)
optimized.design$ofv
optimized.design$rse
`

sessioninfo as below:

R version 3.6.3 (2020-02-29)

Platform: x86_64-w64-mingw32/x64 (64-bit)

Running under: Windows 10 x64 (build 17763)

Matrix products: default

locale:

[1] LC_COLLATE=English_United States.1252

[2] LC_CTYPE=English_United States.1252

[3] LC_MONETARY=English_United States.1252

[4] LC_NUMERIC=C

[5] LC_TIME=English_United States.1252

attached base packages:

[1] stats graphics grDevices utils datasets methods base

other attached packages:

[1] RxODE_0.9.2-0 PopED_0.4.0

loaded via a namespace (and not attached):

[1] Rcpp_1.0.4.6 pillar_1.4.4 compiler_3.6.3

[4] sys_3.2 PreciseSums_0.3 prettyunits_1.0.2

[7] remotes_2.1.1 tools_3.6.3 mvnfast_0.2.5

[10] testthat_2.3.2 digest_0.6.25 pkgbuild_1.0.6

[13] pkgload_1.0.2 tibble_3.0.1 memoise_1.1.0

[16] lifecycle_0.2.0 gtable_0.3.0 pkgconfig_2.0.2

[19] rlang_0.4.5 cli_2.0.2 rstudioapi_0.11

[22] curl_3.3 dplyr_0.8.5 withr_2.1.2

[25] vctrs_0.2.4 desc_1.2.0 fs_1.3.1

[28] devtools_2.3.0 tidyselect_0.2.5 rprojroot_1.3-2

[31] grid_3.6.3 data.table_1.12.8 glue_1.4.0

[34] R6_2.4.1 processx_3.4.0 fansi_0.4.1

[37] sessioninfo_1.1.1 dparser_0.1.8 purrr_0.3.4

[40] ggplot2_3.3.0 callr_3.4.3 magrittr_1.5

[43] units_0.6-3 backports_1.1.4 scales_1.0.0

[46] ps_1.3.0 ellipsis_0.3.0 usethis_1.6.1

[49] assertthat_0.2.1 colorspace_1.4-1 labeling_0.3

[52] lotri_0.2.1 munsell_0.5.0 crayon_1.3.4

commented

Hi

The answer to both of your questions is to use the development version of PopED from GitHub.

All plots created by PopED are ggplot2 objects and can be manipulated as you would any other ggplot created plot.

Best regards,
Andrew

commented