Error message with pmx_nlmixr

[nskoch]

Error message with pmx_nlmixr

[agstudy]

How can I reproduce this error?
Is it a public example?

[nskoch]

It's personnal data previously analyzed with NONMEM (FOCEI and SAEM).

[nskoch]

head(d3)
ID TIME DV GR AMT BSL EVID
1 1 7:30 0 1 27586 0 1
2 1 9:00 732 1 0 0 0
3 1 12:00 152 1 0 0 0
4 1 12:30 0 1 27586 0 1
5 1 14:00 701 1 0 0 0
6 1 16:00 452 1 0 0 0

one.cmt.ka <- function() {
ini({
tka <- 0.1 ; label("log KA")
tcl <- log(5) ; label("log Cl")
tv <- log(20) ; label("log V")
eta.ka ~ 0.1
eta.cl ~ 0.3
eta.v ~ 0.1
add.err <- 10
})
model({
cl <- exp(tcl + eta.cl)
v <- exp(tv + eta.v)
Ka <- exp(tka + eta.ka)
linCmt() ~ add(add.err)
})
}
fit_ka_focei <- nlmixr(one.cmt.ka, d3, est="focei",table = list(cwres=TRUE, npde=TRUE))

print(fit_ka_focei)
-- nlmixr FOCEi (outer: nlminb) fit --------------------------------------------
OBJF AIC BIC Log-likelihood Condition Number
FOCEi 4376.714 5050.512 5077.695 -2518.256 201.2068

-- Time (sec; fit_ka_focei$time): ----------------------------------------------
setup optimize covariance table npde other
elapsed 30.487 5.182 5.182 0.16 2.59 50.729
.../...

ctr <- pmx_nlmixr(fit_ka_focei)
Compiling model...done
Compiling VPC model...done
done (2.50 sec)
Error in pmx_nlmixr(fit_ka_focei) :
Cannot merge nlmixr fit with observation dataset

[agstudy]

Can you show me the output of :

FIT <- as.data.frame(fit_ka_focei)
obs <- as.data.table(nlmixr::getData(fit_ka_focei))
 ## obs <- obs[!(EVID == 1 & MDV == 1)]
 if (any(names(obs) == "EVID")) {
   obs <- obs[EVID == 0 || EVID == 2]
 } else if (any(names(obs) == "MDV")) {
   obs <- obs[MDV == 0]
 }
 if (any(names(obs) == "ID")) {
   obs$ID <- paste(obs$ID)
 }
 ## Merge with DV too
 no_cols <- setdiff(intersect(names(FIT), names(obs)), c("ID", "TIME"))
 obs[, (no_cols) := NULL]
 uID <- unique(FIT$ID)
 obs <- subset(obs, ID %in% uID)
 obs$ID <- factor(obs$ID, levels = levels(fit$ID))
 FIT$ID <- factor(FIT$ID, levels = levels(fit$ID))
 head(obs)
 head(FIT)

[nskoch]

obs$ID <- factor(obs$ID, levels = levels(FIT$ID))
FIT$ID <- factor(FIT$ID, levels = levels(FIT$ID))
head(obs)
Empty data.table (0 rows and 5 cols): ID,TIME,GR,AMT,BSL
head(FIT)
ID TIME DV EVID PRED RES WRES IPRED IRES IWRES CPRED CRES
1 1 2 152 0 0.0000 152.00000 0.016129525 0.0000 152.00000 0.60194368 0.0000 152.0000
2 1 4 701 0 568.3839 132.61614 0.014072601 685.2753 15.72470 0.06227228 564.9302 136.0698
3 1 5 452 0 357.2372 94.76279 0.010055782 463.6817 -11.68171 -0.04626139 344.1230 107.8770
4 1 13 732 0 360.5989 371.40110 0.039411338 468.0528 263.94715 1.04527184 347.2910 384.7090
5 2 1 184 0 0.0000 184.00000 0.019525215 0.0000 184.00000 0.72866866 0.0000 184.0000
6 2 2 69 0 0.0000 69.00000 0.007321955 0.0000 69.00000 0.27325075 0.0000 69.0000
CWRES eta.ka eta.cl eta.v rx1c cl v Ka depot
1 0.6019437 0.006593247 -0.2173521 0.007965366 0.0000 14.25946 6.774591 0.5891999 0.000
2 0.4383556 0.006593247 -0.2173521 0.007965366 685.2751 14.25946 6.774591 0.5891999 15303.912
3 0.3558046 0.006593247 -0.2173521 0.007965366 463.6817 14.25946 6.774591 0.5891999 8490.167
4 1.2648771 0.006593247 -0.2173521 0.007965366 468.0530 14.25946 6.774591 0.5891999 8566.344
5 0.7286687 -0.010607003 0.2939231 -0.006954881 0.0000 23.77645 6.674263 0.5791522 0.000
6 0.2732507 -0.010607003 0.2939231 -0.006954881 0.0000 23.77645 6.674263 0.5791522 0.000
central EPRED ERES NPDE
1 0.000 7.28364 144.7164 0.8479605
2 4642.459 831.64152 -130.6415 0.2540297
3 3141.254 569.40459 -117.4046 0.4447049
4 3170.868 412.81680 319.1832 1.3796618
5 0.000 27.64481 156.3552 -0.4891047
6 0.000 -200.28132 269.2813 0.9187177

[agstudy]

It looks that the obs is empty due to a certain filter , can you run this please ? :

obs <- as.data.table(nlmixr::getData(fit_ka_focei))
head(obs)
 if (any(names(obs) == "EVID")) {
   obs <- obs[EVID == 0 || EVID == 2]
 } else if (any(names(obs) == "MDV")) {
   obs <- obs[MDV == 0]
 }
head(obs)
 uID <- unique(FIT$ID)
 obs <- subset(obs, ID %in% uID)
 head(obs) 

[nskoch]

Here it is

obs <- as.data.table(nlmixr::getData(fit_ka_focei))
head(obs)
ID TIME DV GR AMT BSL EVID
1: 1 7:30 0 1 27586 0 1
2: 1 9:00 732 1 0 0 0
3: 1 12:00 152 1 0 0 0
4: 1 12:30 0 1 27586 0 1
5: 1 14:00 701 1 0 0 0
6: 1 16:00 452 1 0 0 0
if (any(names(obs) == "EVID")) {

• obs <- obs[EVID == 0 || EVID == 2]
• } else if (any(names(obs) == "MDV")) {
• obs <- obs[MDV == 0]
• }

head(obs)
Empty data.table (0 rows and 7 cols): ID,TIME,DV,GR,AMT,BSL...
uID <- unique(FIT$ID)
Error in unique(FIT$ID) : object 'FIT' not found
obs <- subset(obs, ID %in% uID)
Error in ID %in% uID : object 'uID' not found
head(obs)
Empty data.table (0 rows and 7 cols): ID,TIME,DV,GR,AMT,BSL...

[agstudy]

it should be fixed now. can reinstall ggPMX please.

[nskoch]

I still have the same problem:

ctr <- pmx_nlmixr(fit_ka_focei)Compiling VPC model...donedone (0.89 sec)Error in pmx_nlmixr(fit_ka_focei) :

Cannot merge nlmixr fit with observation dataset Le jeu. 19 déc. 2019 à 10:38, agstudy <[email protected]> a écrit :

…

it should be fixed now. can reinstall ggPMX please. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#41?email_source=notifications&email_token=AKBGW6G5VTACZEF5F54WL2TQZM6KPA5CNFSM4J3XU5XKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHJASTA#issuecomment-567413068>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKBGW6BA7XF6VJW5GCBP7FDQZM6KPANCNFSM4J3XU5XA> .

-- Pharmacology & Pharmacokinetic Consultant Professor of Medicine, University of Aix-Marseille Head of the Department of Clinical Pharmacology, Ste Marguerite Hospital Office: [phone] Mobile: [phone]

[agstudy]

What are the values of EVID in your observation data set?

…

On Fri, 20 Dec 2019 at 18:16, nskoch ***@***.***> wrote: I still have the same problem: > ctr <- pmx_nlmixr(fit_ka_focei)Compiling VPC model...donedone (0.89 sec)Error in pmx_nlmixr(fit_ka_focei) : Cannot merge nlmixr fit with observation dataset Le jeu. 19 déc. 2019 à 10:38, agstudy ***@***.***> a écrit : > it should be fixed now. can reinstall ggPMX please. > > — > You are receiving this because you authored the thread. > Reply to this email directly, view it on GitHub > < #41?email_source=notifications&email_token=AKBGW6G5VTACZEF5F54WL2TQZM6KPA5CNFSM4J3XU5XKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHJASTA#issuecomment-567413068 >, > or unsubscribe > < https://github.com/notifications/unsubscribe-auth/AKBGW6BA7XF6VJW5GCBP7FDQZM6KPANCNFSM4J3XU5XA > > . > -- Pharmacology & Pharmacokinetic Consultant Professor of Medicine, University of Aix-Marseille Head of the Department of Clinical Pharmacology, Ste Marguerite Hospital Office: [phone] Mobile: [phone] — You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <#41?email_source=notifications&email_token=AAN75R2C3R7XDE4CFS4QXKLQZT4XLA5CNFSM4J3XU5XKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHNRNDQ#issuecomment-568006286>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAN75R5ETD72NICZB5KELRLQZT4XLANCNFSM4J3XU5XA> .

[nskoch]

1 if AMT, 0 otherwise

…

Le 20 déc. 2019 à 19:57, agstudy ***@***.***> a écrit : What are the values of EVID in your observation data set? On Fri, 20 Dec 2019 at 18:16, nskoch ***@***.***> wrote: > I still have the same problem: > > > ctr <- pmx_nlmixr(fit_ka_focei)Compiling VPC model...donedone (0.89 > sec)Error in pmx_nlmixr(fit_ka_focei) : > Cannot merge nlmixr fit with observation dataset > > > Le jeu. 19 déc. 2019 à 10:38, agstudy ***@***.***> a écrit : > > > it should be fixed now. can reinstall ggPMX please. > > > > — > > You are receiving this because you authored the thread. > > Reply to this email directly, view it on GitHub > > < > #41?email_source=notifications&email_token=AKBGW6G5VTACZEF5F54WL2TQZM6KPA5CNFSM4J3XU5XKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHJASTA#issuecomment-567413068 > >, > > or unsubscribe > > < > https://github.com/notifications/unsubscribe-auth/AKBGW6BA7XF6VJW5GCBP7FDQZM6KPANCNFSM4J3XU5XA > > > > . > > > > > -- > Pharmacology & Pharmacokinetic Consultant > Professor of Medicine, University of Aix-Marseille > Head of the Department of Clinical Pharmacology, Ste Marguerite Hospital > Office: [phone] > Mobile: [phone] > > — > You are receiving this because you were assigned. > Reply to this email directly, view it on GitHub > <#41?email_source=notifications&email_token=AAN75R2C3R7XDE4CFS4QXKLQZT4XLA5CNFSM4J3XU5XKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHNRNDQ#issuecomment-568006286>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAN75R5ETD72NICZB5KELRLQZT4XLANCNFSM4J3XU5XA> > . > — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[agstudy]

amazing. very hard to debug without a reproducible data aset.

It looks that the obs is empty due to a certain filter , can you run this please ? :
Can you rerun this part to see what message do you have please?

obs <- as.data.table(nlmixr::getData(fit_ka_focei))
head(obs)
 if (any(names(obs) == "EVID")) {
   obs <- obs[EVID == 0 | EVID == 2]
 } else if (any(names(obs) == "MDV")) {
   obs <- obs[MDV == 0]
 }
head(obs)
 uID <- unique(FIT$ID)
 obs <- subset(obs, ID %in% uID)
 head(obs) 

[nskoch]

Thanks for your help. Here it is:

obs <- as.data.table(nlmixr::getData(fit_ka_focei))> head(obs) ID TIME DV GR AMT BSL EVID

1: 1 7:30 0 1 27586 0 1 2: 1 9:00 732 1 0 0 0 3: 1 12:00 152 1 0 0 0 4: 1 12:30 0 1 27586 0 1 5: 1 14:00 701 1 0 0 0 6: 1 16:00 452 1 0 0 0> if (any(names(obs) == "EVID")) {+ obs <- obs[EVID == 0 | EVID == 2]+ } else if (any(names(obs) == "MDV")) {+ obs <- obs[MDV == 0]+ }> head(obs) ID TIME DV GR AMT BSL EVID 1: 1 9:00 732 1 0 0 0 2: 1 12:00 152 1 0 0 0 3: 1 14:00 701 1 0 0 0 4: 1 16:00 452 1 0 0 0 5: 2 8:00 493 1 0 14 0 6: 2 10:00 184 1 0 14 0> uID <- unique(FIT$ID)Error in unique(FIT$ID) : object 'FIT' not found> uID <- unique(fit_ka_focei$ID)> obs <- subset(obs, ID %in% uID)> head(obs) ID TIME DV GR AMT BSL EVID 1: 1 9:00 732 1 0 0 0 2: 1 12:00 152 1 0 0 0 3: 1 14:00 701 1 0 0 0 4: 1 16:00 452 1 0 0 0 5: 2 8:00 493 1 0 14 0 6: 2 10:00 184 1 0 14 0> # PMX code> ctr <- pmx_nlmixr(fit_ka_focei)Compiling model...doneCompiling VPC model...donedone (2.25 sec)Error in pmx_nlmixr(fit_ka_focei) : Cannot merge nlmixr fit with observation dataset Le lun. 23 déc. 2019 à 01:21, agstudy <[email protected]> a écrit :

…

amazing. very hard to debug without a reproducible data aset. It looks that the obs is empty due to a certain filter , can you run this please ? : Can you rerun this part to see what message do you have please? obs <- as.data.table(nlmixr::getData(fit_ka_focei)) head(obs) if (any(names(obs) == "EVID")) { obs <- obs[EVID == 0 | EVID == 2] } else if (any(names(obs) == "MDV")) { obs <- obs[MDV == 0] } head(obs) uID <- unique(FIT$ID) obs <- subset(obs, ID %in% uID) head(obs) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#41?email_source=notifications&email_token=AKBGW6DMFV3JYLWIGSU6WVDQ2AABZA5CNFSM4J3XU5XKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHP5CVQ#issuecomment-568316246>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKBGW6AQ6YAEUJARARPD4ZTQ2AABZANCNFSM4J3XU5XA> .

-- Pharmacology & Pharmacokinetic Consultant Professor of Medicine, University of Aix-Marseille Head of the Department of Clinical Pharmacology, Ste Marguerite Hospital Office: [phone] Mobile: [phone]

[nskoch]

Here is my full code and the data file. Hope that helps. Best regards Nicolas d3 <- read.csv(paste(wp, "data.csv", sep="/") ) head(d3) names(d3)[1]<-"ID" d3$EVID <- ifelse(d3$AMT>0,1,0) head(d3) one.cmt.ka <- function() { ini({ tka <- 0.1 ; label("log KA") tcl <- log(5) ; label("log Cl") tv <- log(20) ; label("log V") eta.ka ~ 0.1 eta.cl ~ 0.3 eta.v ~ 0.1 add.err <- 10 }) model({ cl <- exp(tcl + eta.cl) v <- exp(tv + eta.v) Ka <- exp(tka + eta.ka) linCmt() ~ add(add.err) }) } library(ggPMX) library(xgxr) fit_ka_focei <- nlmixr(one.cmt.ka, d3, est="focei", table = list(cwres=TRUE, npde=TRUE)) Le lun. 23 déc. 2019 à 01:21, agstudy <[email protected]> a écrit :

…

amazing. very hard to debug without a reproducible data aset. It looks that the obs is empty due to a certain filter , can you run this please ? : Can you rerun this part to see what message do you have please? obs <- as.data.table(nlmixr::getData(fit_ka_focei)) head(obs) if (any(names(obs) == "EVID")) { obs <- obs[EVID == 0 | EVID == 2] } else if (any(names(obs) == "MDV")) { obs <- obs[MDV == 0] } head(obs) uID <- unique(FIT$ID) obs <- subset(obs, ID %in% uID) head(obs) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#41?email_source=notifications&email_token=AKBGW6DMFV3JYLWIGSU6WVDQ2AABZA5CNFSM4J3XU5XKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHP5CVQ#issuecomment-568316246>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKBGW6AQ6YAEUJARARPD4ZTQ2AABZANCNFSM4J3XU5XA> .

-- Pharmacology & Pharmacokinetic Consultant Professor of Medicine, University of Aix-Marseille Head of the Department of Clinical Pharmacology, Ste Marguerite Hospital Office: [phone] Mobile: [phone]

[agstudy]

thank you for sending the code. But I don't see the data file in the attachment. You can send it by email (maintainer email)

[agstudy]

close issue since no data was provided.