Re:Re: Parent and metabolite model-residual error model and L2
Dear Nick and Ceon,
Thank you very much for your immediate replies.
Best regards,
Xipei
--
Xipei Wang, Ph.D. student
Department of Pharmaceutics, School of Pharmaceutical Sciences,
Peking University Health Science Center, Beijing,
China
Email: [email protected]
At 2011-09-02 16:07:37，"Nick Holford" <[email protected]> wrote:
>Xipei,
>
>For Question 1 -- you need to specify both L2 and SIGMA(BLOCK).
>
>I think the answer to your Question 2 is well described in the online
>help for NONMEM.
>Perhaps you are not aware that you can find help for NM-TRAN/NONMEM as
>part of a normal NONMEM installation. You need to look for the
>html\index.html file, open it in an internet browser (save the URL as a
>favourite!), then search e.g. click on L for information about the L2
>data item.
>
>For Question 3 -- I think L2 is only relevant for observation records
>(MDV=0 or EVID=0). Therefore I think you can use any value you look for
>L2 on a dosing record. A sensible value might be '.' which is meaningful
>to human readers to indicate that this is a missing value (which, in
>this case, does not matter) . You could also use '42' if you like
>hitchhiking.
>
>Note that there is no need to use POSTHOC with METHOD=1. The CONDITIONAL
>method always returns POSTHOC values in the output table.
>
>Best wishes,
>
>Nick
>
>DISCUSSION:
>L2 labels level-two (L2) data items. The level two data item is
>optional.
>
>Recall that the ID data item is used to group together the data
>records containing observations which have the same realization of the
>level-one random effects (etas). Similarly, the L2 data item is used
>to group together the data records containing observations which have
>the same realization of the level-two random effects (epsilons). The
>observations of such a group is called a level-two observation. The
>group itself is called a level-two (L2) record. Data records of an L2
>record must be contiguous (and contained within the same individual
>record). By default, level-two observations are treated as being sta-
>tistically independent multivariate observations. (However, within a
>level-one observation, the level-two random effects can be made to be
>correlated between level-two observations.)
>
>Here is an example of a fragment of a data set using L2 data items.
>There are two types of observations, designated by the two different
>values of the user data item TYPE. Note that the L2 data items are
>the same for both of the observations at TIME=2, and they are also the
>same for both of the observations at TIME=112.5.
>
>ID TIME AMT APGR WT DV TYPE L2
>1 0. 25.0 1.4 7 . . 1
>1 2.0 . 1.4 7 6.0 1 1
>1 2.0 . 1.4 7 17.3 2 1
>1 12.5 3.5 1.4 7 . 2 2
>1 24.5 3.5 1.4 7 . 2 2
>1 37.0 3.5 1.4 7 . 2 2
>1 48.0 3.5 1.4 7 . 2 2
>1 60.5 3.5 1.4 7 . 2 2
>1 72.5 3.5 1.4 7 . 2 2
>1 85.3 3.5 1.4 7 . 2 2
>1 96.5 3.5 1.4 7 . 2 2
>1 108.5 3.5 1.4 7 . 2 2
>1 112.5 . 1.4 7 8.0 1 2
>1 112.5 . 1.4 7 31.0 2 2
>
>
Quoted reply history
>On 2/09/2011 6:31 p.m., wangxipei wrote:
>> Dear NONMEM users,
>>
>> I am building a popPK model for a parent drug and its metabolite (rich
>> data,single dose). I want to estimate the correlation between the
>> residual errors of parent drug and its metabolite, because their
>> measurements were from a same blood sample. (My code and part of data
>> are shown as follows.)
>> Question 1: Should I use $SIGMA BLOCK (2) only, or should I use both
>> L2 and $SIGMA BLOCK(2)?
>> Question 2: I am not sure about the format of L2 data item, so I show
>> the first two individuals.
>> Within one individual, the observations of parent and its metabolite
>> at the same time point have the same L2 value, but different from
>> other time points. Is it correct?
>> Question 3: Does the L2 value of the dose event affect the estimation?
>> (I think it ! does not.)
>>
>>
>> ----------------------------------- NONMEM code
>> ------------------------------------------------------------------------------------------
>> $INPUT C ID TIME DV AMT CMT EVID L2
>> $DATA d.CSV IGNORE=@
>> $SUBROUTINES ADVAN6 TOL=6
>> $MODEL NCOMP=7 COMP(DEFDEP, DEFDOSE) COMP(CENTRAL, DEFOBS)
>> COMP(PERIPH) COMP(META) COMP(META2)
>> COMP(TRANSIT1) COMP(TRANSIT2) ; 2-COMP for parent drug, 2-comp for its
>> metabolite, 2-transit compartments to connect them.
>> $PK
>> CL=TVCL*EXP(ETA(1))
>> ...
>> $DES
>> ...
>> $ERROR
>> DEL=0
>> IF (F.EQ.0) DEL=1
>> W=F
>> IPRED=F
>> IRES=DV-IPRED
>> IWRES=IRES/(W+DEL)
>> IF(CMT.EQ.2) Y=F + F*ERR(1)
>> IF(CMT.EQ.4) Y=F + F*ERR(2) + ERR(3)!
>> ...
>> $SIGMA BLOCK(2)
>> .2 ;parent RV
>> .1 .17 ;metabolite RV
>> $SIGMA BLOCK(1) 0 FIX;[A]
>> $EST PRINT=5 MAX=9999 SIG=3 METH=1 POSTHOC MSFO=run1.MSF
>>
>>
>> --------------individual data
>> --------------------------------------------------------------
>> compartment 2 is the central compartment of parent drug, and
>> compartment 4 is the central comp. of metabolite.
>>
>> ID TIME DV AMT CMT EVID L2 1 0 . 100 1 1 0 1 0.33 1.096 . 2 0 1 1 0.66
>> 0.623 . 2 0 2 1 0.66 0.130 . 4 0 2 1 1 0.329 . 2 0 3 1 1 0.407 . 4 0 3
>> 1 1.5 0.139 . 2 0 4 1 1.5 0.541 . 4 0 4 1 2 0.073 . 2 0 5 1 2 0.552 .
>> 4 0 5 1 2.5 0.022 . 2 0 6 1 2.5 0.465 . 4 0 6 1 3 0.076 . 2 0 7 1 3
>> 0.572 . 4 0 7 1 3.5 0.479 . 4 0 8 1 4 0.480 . 4 0 9 1 6 0.160 . 4 0 10
>> 1 8 0.252 . 4 0 11 1 12 0.117 . 4 0 12 1 24 0.018 . 4 0 13 2 0 . 100 1
>> 1 0 2 0.33 0.030 . 2 0 1 2 1 0.034 . 2 0 3 2 1.5 0.049 . 2 0 4 2 2
>> 0.053 . 2 0 5 2 2.5 0.299 . 2 0 6 2 2.5 0.033 . 4 0 6 2 3 0.344 . 2 0
>> 7 2 3 0.118 . 4 0 7 2 3.5 0.263 . 2 0 8 2 3.5 0.293 . 4 0 8
>>
>>
>>
>> Many thanks in advance for any comments!
>> Xipei
>>
>> --
>> Xipei Wang, Ph.D. student
>> Department of Pharmaceutics, School of Pharmaceutical Sciences,
>> Peking University Health Science Center, Beijing,
>> China
>> Email: [email protected]
>>
>>
>
>--
>Nick Holford, Professor Clinical Pharmacology
>Dept Pharmacology& Clinical Pharmacology
>University of Auckland,85 Park Rd,Private Bag 92019,Auckland,New Zealand
>tel:+64(9)923-6730 fax:+64(9)373-7090 mobile:+64(21)46 23 53
>email: [email protected]
> http://www.fmhs.auckland.ac.nz/sms/pharmacology/holford
>