NONMEM Users Network Archive

Hosted by Cognigen

RE: [NMusers] BAYES method and initial estimate bounds

From: Bob Leary <Bob.Leary_at_certara.com>
Date: Tue, 22 Dec 2015 20:56:48 +0000

1.34+154 is often on many systems used to indicate an improbably large floa=
ting point number that is likely an error of some sort. It is the sqrt of t=
he largest representable real*8 value, 1.797+308


________________________________
From: owner-nmusers_at_globomaxnm.com [owner-nmusers_at_globomaxnm.com] on behalf=
 of Bauer, Robert [Robert.Bauer_at_iconplc.com]
Sent: Tuesday, December 22, 2015 1:40 PM
To: Bill Gillespie
Cc: nmusers_at_globomaxnm.com
Subject: RE: [NMusers] BAYES method and initial estimate bounds

Bill:
Yes, and when that happens, the execution path is difficult to anticipate.

Robert J. Bauer, Ph.D.
Vice President, Pharmacometrics R&D
ICON Early Phase
Office: (215) 616-6428
Mobile: (925) 286-0769
Robert.Bauer_at_iconplc.com<mailto:Robert.Bauer_at_iconplc.com>
www.iconplc.com<http://www.iconplc.com>

From: Bill Gillespie [mailto:bgillespiester_at_gmail.com]
Sent: Tuesday, December 22, 2015 11:25 AM
To: Bauer, Robert
Cc: nmusers_at_globomaxnm.com
Subject: Re: [NMusers] BAYES method and initial estimate bounds

Hi Bob,

Thanks for the reply. On investigating further I see this happens when NONM=
EM reports a THETA value of 1.34+154 in the PRDERR file. Would I be correct=
 in thinking that is the result of a floating point error somewhere in the =
sampling process?

Thanks,
Bill

On Tue, Dec 22, 2015 at 12:40 PM, Bauer, Robert <Robert.Bauer_at_iconplc.com<m=
ailto:Robert.Bauer_at_iconplc.com>> wrote:
Bill:
I just tested this on a simple script (see below), and the boundaries are i=
mposed in BAYES analysis for me.

$PROB RUN# Example 1 (from samp5l)
$INPUT C SET ID JID TIME DV=CONC AMT=DOSE RATE EVID MDV CMT CLX V1X QX=
 V2X SDIX SDSX
$DATA example1.csv IGNORE=C
$SUBROUTINES ADVAN3 TRANS4
$PK
MU_1=THETA(1)
MU_2=THETA(2)
MU_3=THETA(3)
MU_4=THETA(4)
IF (THETA(1)<1.63.OR.THETA(1)>1.8) WRITE(50,*) THETA(1)
CL=DEXP(THETA(1)+ETA(1))
V1=DEXP(THETA(2)+ETA(2))
Q=DEXP(MU_3+ETA(3))
V2=DEXP(MU_4+ETA(4))
S1=V1
$ERROR
Y = F + F*EPS(1)
$THETA (1.63, 1.67,1.8) 2.0 2.0 2.0
$OMEGA BLOCK(4) VALUES(0.15,0.01)
$SIGMA (0.6 )
$PRIOR NWPRI
$THETAP (2.0 FIX) (2.0 FIX) (2.0 FIX) (2.0 FIX)
$THETAPV BLOCK(4) FIX VALUES(10000,0.0)
$OMEGAP BLOCK(4) FIX VALUES(0.2,0.0)
$OMEGAPD (4 FIX)
$EST METHOD=BAYES INTERACTION NITER=1000 PRINT=100 CTYPE=3
$COV MATRIX=R PRINT=E UNCONDITIONAL

Robert J. Bauer, Ph.D.
Vice President, Pharmacometrics R&D
ICON Early Phase
Office: (215) 616-6428<tel:%28215%29%20616-6428>
Mobile: (925) 286-0769<tel:%28925%29%20286-0769>
Robert.Bauer_at_iconplc.com<mailto:Robert.Bauer_at_iconplc.com>
www.iconplc.com<http://www.iconplc.com>

From: owner-nmusers_at_globomaxnm.com<mailto:owner-nmusers_at_globomaxnm.com> [ma=
ilto:owner-nmusers_at_globomaxnm.com<mailto:owner-nmusers_at_globomaxnm.com>] On =
Behalf Of Bill Gillespie
Sent: Tuesday, December 22, 2015 8:14 AM
To: nmusers_at_globomaxnm.com<mailto:nmusers_at_globomaxnm.com>
Subject: [NMusers] BAYES method and initial estimate bounds

Hi all,

If bounds are specified in the $THETA statements for initial estimates, wha=
t, if any, effect do they have when the BAYES method is used?

I initially thought they might work in conjunction with the prior specifica=
tions to implement truncated priors. However that does not seem to be true =
because statements like "IF(ABS(THETA(1)) > 10) EXIT 1 14" sometimes get tr=
ipped even when THETA(1) is bounded between -10 and 10 in the $THETA state=
ment.

Happy Holidays,
Bill


ICON plc made the following annotations.
---------------------------------------------------------------------------=
---
This e-mail transmission may contain confidential or legally privileged inf=
ormation that is intended only for the individual or entity named in the e-=
mail address. If you are not the intended recipient, you are hereby notifie=
d that any disclosure, copying, distribution, or reliance upon the contents=
 of this e-mail is strictly prohibited. If you have received this e-mail tr=
ansmission in error, please reply to the sender, so that ICON plc can arran=
ge for proper delivery, and then please delete the message.

Thank You,

ICON plc
South County Business Park
Leopardstown
Dublin 18
Ireland
Registered number: 145835 NOTICE: The information contained in this=
 electronic mail message is intended only for the personal and confidential=
  use of the designated recipient(s) named above. This message may be a=
n attorney-client communication, may be protected by the work product =
doctrine, and may be subject to a protective order. As such, this message i=
s privileged and confidential. If the reader of this message is not th=
e intended recipient or an agent responsible for delivering it to the =
intended recipient, you are hereby notified that you have received this mes=
sage in error and that any review, dissemination, distribution, or cop=
ying of this message is strictly prohibited. If you have received this =

mail and destroy any and all copies of this message in your possession=
 (whether hard copies or electronically stored copies). Thank you. buS=
p9xeMeKEbrUze


Received on Tue Dec 22 2015 - 15:56:48 EST

The NONMEM Users Network is maintained by ICON plc. Requests to subscribe to the network should be sent to: nmusers-request@iconplc.com. Once subscribed, you may contribute to the discussion by emailing: nmusers@globomaxnm.com.