Re: [NMusers] bug/feature in $AES

Thu, 27 Sep 2007 06:57:07 -0700 

Hi Nick,
It would help if you prepare a cleaner example that demonstrate the problem. What exactly are you trying to achieve (A(?)=?). A(1) and A(2) are not initialized in the code below, so they are zeros as they should. Is this about A3, A4? 


If yes, the problem could be in communication between $AES and $PK. Is 
it allowed to use A() in the PK block?

>    GBASE=A(3)/VGLU ; GLUCOSE INITIAL
>    IBASE=A(4)/VINS ; INSULIN INITIAL
From help:

"   Right-hand quantities (of PK block can contain - LG) in assignment 
statement and in conditions:

     Data item labels specified on the $INPUT statement.
     THETA(n).
     ETA(n) (Used if the data are population.)
     PK-defined items that appeared earlier as left-hand quantities.



I think, PK and blocks below can communicate via COMMONs but you need to 
take special action (via COMRES) to make them available, and they have a 
1-record delay (next record knows about COM() of the previous record). I 
could be wrong on that, did it just once, and it was related to ERROR, 
not AES, but we put extra record with the same time, just to pass 
variable to PK


Leonid


Nick Holford wrote:

Hi,
[Second attempt - there was an error in the copy of the control stream in our 
first attempt]
Carl Kirkpatrick and I have been trying to use $AES to compute initial values 
for use with $DES.
We can get $AES to produce the correct equilibrium values at times after time=0 
but cannot access these values at time=0 or at time=0.001. Because the 
equilibrium (i.e. initial condition) values are not available at time=0 we 
cannot use the NMVI method for initalizing the amounts in the compartments 
using A_0(1)=GBASE and A_0(2)=IBASE.

We have inserted dosing records using the old NMV bioavailability trick to try 
and initialize the compartments at time=0.002. This almost gets the correct 
solution at time=0.002 in both compartments.

Can anyone explain why the algebraic equation results are not available at 
time=0 and time=0.001?

Nick and Carl

--- table file output -----                           GLU=A(1)        INS=A(2)
ID      TIME    AMT     CMT     F1      F2      GBASE   GLU     IBASE   INS
1       0       0       1       0       0       0       0       0       0
1       0       0       2       0       0       0       0       0       0
1       0.001   0       1       0       0       0       0.0009  0       0.0024
1       0.002   1       1       187.14  527.49  4.6785  4.6805  13.187  0.0049
1       0.002   1       2       187.14  527.49  4.6785  4.6805  13.187  13.192
1       100     0       1       0       0       4.6785  4.6785  13.187  13.187
1       100     0       2       0       0       4.6785  4.6785  13.187  13.187
---- Control stream --------
$PROB GLUCOSE AND INSULIN INITIAL VALUE
$DATA test.csv
$INPUT ID TIME CMT AMT DV
$SIM (200070927) ONLYSIM NSUB=1

$THETA
40 ; POP_RGLU MMOL/H/70KG

40 ; POP_VGLU L/70KG 
4 ; POP_CLGLU L/H/70KG

1 ; POP_EMXGLU
10 ; POP_C50GLU MMOL/L
100 ; POP_RINS NMOL/H/70KG
40 ; POP_VINS L/70KG
10 ; POP_CLINS L/H/70KG
2 ; POP_EMXINS
10 ; POP_C50INS NMOL/L

$OMEGA
0 FIX ;PPV_RGLU
0 FIX ;PPV_RINS

 
$SIGMA

0.1  ;G_EXP_RUV
1    ;G_ADD_RUV
0.1  ;I_EXP_RUV
1    ;I_ADD_RUV

$SUBR ADVAN9 TOL=3

$MODEL
   COMP (GLUCOSE)
   COMP (INSULIN)
   COMP (GLU EQUILIBRIUM)
   COMP (INS EQUILIBRIUM)

   
$PK
"FIRST 
" COMMON/PRCOMG/IDUM1,IDUM2,IMAX 
" INTEGER IDUM1,IDUM2,IMAX 
" IMAX=1000000


   ; GLUCOSE
   RGLU=THETA(1)*EXP(ETA(1))

   VGLU=THETA(2) 
   CLGLU=THETA(3)

   EMXGLU=THETA(4)
   C50GLU=THETA(5)

   
   ;INSULIN   
   RINS=THETA(6)*EXP(ETA(2))

   VINS=THETA(7)
   CLINS=THETA(8)
   EMXINS=THETA(9)
   C50INS=THETA(10)

   GBASE=A(3)/VGLU ; GLUCOSE INITIAL
   IBASE=A(4)/VINS ; INSULIN INITIAL
   IF (AMT.GT.0) THEN
      F1=GBASE*VGLU
      F2=IBASE*VINS
   ELSE
      F1=0
      F2=0
   ENDIF
   S1=VGLU
   S2=VINS

$AESINITIAL
   INIT=0

   A(3)=5 ; MMOL/L 
   A(4)=10 ; NMOL/L

$AES
   EGLU=A(3)/VGLU
   EINS=A(4)/VINS
   EGEFF=EMXGLU*EGLU/(C50GLU+EGLU)
   EIEFF=EMXINS*EINS/(C50INS+EINS)
   E(3)=RGLU - CLGLU*(1+EIEFF)*EGLU ; GLUCOSE EQUILIBRIUM
   E(4)=RINS*(1+EGEFF) - CLINS*EINS ; INSULIN EQUILIBRIUM

$DES                                          
   DGLU=A(1)/VGLU                          
   DINS=A(2)/VINS  
   DGEFF=EMXGLU*DGLU/(C50GLU+DGLU)

   DIEFF=EMXINS*DINS/(C50INS+DINS)
   DADT(1)=RGLU - CLGLU*(1+DIEFF)*DGLU
   DADT(2)=RINS*(1+DGEFF)- CLINS*DINS

$ERROR
GLU=A(1)/VGLU
INS=A(2)/VINS

IF (CMT.EQ.1)THEN 
  Y=GLU*(1+ERR(1))+ERR(2)

ENDIF

  
IF (CMT.EQ.2)THEN

  Y=INS*(1+ERR(3))+ERR(4)
ENDIF


$TABLE ID TIME AMT CMT F1 F2 GBASE GLU IBASE INS  
NOAPPEND ONEHEADER NOPRINT FILE=test.fit

---- test.dat --------------------
#id,time,cmt,amount,dv
1,0,1,.,.
1,0,2,.,.
1,0.001,1,.,.
1,0.002,1,1,.
1,0.002,2,1,.
1,100,1,.,.
1,100,2,.,.

--
Nick Holford, Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
[EMAIL PROTECTED] tel:+64(9)373-7599x86730 fax:+64(9)373-7090
www.health.auckland.ac.nz/pharmacology/staff/nholford
























					Reply via email to