Dear NM users
I have a problem I need your help, please.
I’m trying to fit the PD model where I used PK parameters as variables in
the dataset. The model file is below, it runs BUT it seems the parameter
does not affect prediction, and hence the model terminated due to zero
gradient (see the output).
I will appreciate your help.
Thanks,
Ali
0ITERATION NO.: 0 OBJECTIVE VALUE: 1255.70665291810 NO. OF
FUNC. EVALS.: 5
CUMULATIVE NO. OF FUNC. EVALS.: 5
NPARAMETR: 1.0000E-03 1.0500E+00 1.0000E-03 9.0000E-02
PARAMETER: 1.0000E-01 1.0000E-01 1.0000E-01 1.0000E-01
GRADIENT: 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00
0ITERATION NO.: 2 OBJECTIVE VALUE: 1255.70665291810 NO. OF
FUNC. EVALS.: 9
CUMULATIVE NO. OF FUNC. EVALS.: 14
NPARAMETR: 1.0000E-03 1.0500E+00 1.0000E-03 9.0000E-02
PARAMETER: 1.0000E-01 1.0000E-01 1.0000E-01 1.0000E-01
GRADIENT: 0.0000E+00 0.0000E+00 0.0000E+00 0.0000E+00
Here is how I wrote the model
$INPUT ID TIME TAD AMT AMT2 DV EVID MDV LLOQ
IVMAX IV2 IKA IQ IV3 IKM GROUP DOSELEV DOSE
DOSED1 DOSED2 DOSED3 DOSED4 STUDYN TRT PBASE
$DATA data_pd_08082023.csv IGNORE=@ IGNORE(GROUP.EQ.1)
$SUBROUTINE ADVAN9 TOL=3
$MODEL COMP=(ABS DEFDOSE)
COMP=(CENTRAL DEFOBS)
COMP=(PERIPH1)
COMP=(PALIVE)
COMP=(PDEAD)
$PK
BSV_EC50 = ETA(1)
VMAX = IVMAX
V2 = IV2
KA = IKA
Q = IQ
V3 = IV3
KM = IKM
; scale predictions based on dose (mg) and Cp (ng/mL)
S2 = V2
;DEFINE PD RARAMETERS
IF(GROUP.EQ.1) GR = THETA(1) ; Growth rate of parasite (to be estimated
from control group only)
EC50 = THETA(2)*EXP(BSV_EC50)
HILL = THETA(3)
CLPAR = THETA(4) ; parasite clearance
EMAX = 1/48*3.54 + 0.048
$DES
CC = A(2)/V2
DADT(1) = -KA*A(1)
DADT(2) = KA*A(1) - VMAX*CC/(KM+CC) - (Q/V2)*A(2) + (Q/V3)*A(3)
DADT(3) = (Q/V2)*A(2) - (Q/V3)*A(3)
;;PD
IF(CC.LE.0) CC=0.000001
DEN = CC**HILL + EC50**HILL
KILL = EMAX*(CC**HILL)/DEN
DADT(4) = (GR-KILL)*A(4)
DADT(5) = KILL*A(4) - CLPAR*A(5)
$ERROR
;OBSERVED PARASITEMIA
PALIVE = A(4)
PDEAD = A(5)
POBS = PALIVE + PDEAD
IPRED = POBS
Y = IPRED + ERR(1)
;PD
$THETA 0.001 ; 1 GR
$THETA 1.05; 2 EC50
$THETA 2 FIX ; 3 HILL
$THETA 0.001 ; 4 CLPAR
;
$OMEGA 0.09 ; IIVBSV_EC50
$SIGMA 1 FIX
; Use conditional estimation with interaction
$ESTIMATION METHOD=1 INTER MAXEVAL=9999 PRINT=5 MSFO=run400.msf
