Re: [NMusers] $PRIOR with normal for OMEGA?

From: Mark Sale <msale_at_nuventra.com>
Date: Fri, 11 Nov 2016 14:24:16 +0000

thanks Mats,

  I've never had much luck getting TNPRI to work, pretty limited examples i=
n the docs. I actually not not want to run NONMEM with normal OMEGA PRIOR, =
just want to validate some other work where I calculate the PRIOR penalty f=
rom a normal, and thought this might be a practical way to do it.

I'll look into the TNPRI (didn't realize that stood for triple normal, but =
that makes sense).


thanks

Mark



Mark Sale M.D.
Vice President, Modeling and Simulation
Nuventra Pharma Sciences, Inc.
2525 Meridian Parkway, Suite 280
Durham, NC 27713
Phone (919)-973-0383
msale_at_nuventra.com<msale_at_kinetigen.com>
CONFIDENTIALITY NOTICE The information in this transmittal (including attac=
hments, if any) may be privileged and confidential and is intended only for=
 the recipient(s) listed above. Any review, use, disclosure, distribution o=
r copying of this transmittal, in any form, is prohibited except by or on b=
ehalf of the intended recipient(s). If you have received this transmittal i=
n error, please notify me immediately by reply email and destroy all copies=
 of the transmittal.


________________________________
From: Mats Karlsson <Mats.Karlsson_at_farmbio.uu.se>
Sent: Thursday, November 10, 2016 11:36:13 PM
To: Gewitz, Andrew; Mark Sale
Cc: nmusers_at_globomaxnm.com
Subject: RE: [NMusers] $PRIOR with normal for OMEGA?

Hi Mark,

As I'm sure you know NONMEM has the TNPRI functionality for that. If you wa=
nt to use it without the TNPRI functionality, you can estimate OMEGA as THE=
TA and use a multivariate normal for that. If you have off-diagonal element=
s, you may want to do a Cholesky transformation (you can get that automatic=
ally in PsN).

As Andy writes there are pros and cons with different priors. While IW has =
a better shape to its prior for variances, it is problematic that there is =
no correlation between the typical value estimates and their variances.

Best regards,
Mats


Mats Karlsson, PhD
Professor of Pharmacometrics

Dept of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
75124 Uppsala

Phone: +46 18 4714105
Fax + 46 18 4714003
www.farmbio.uu.se/research/researchgroups/pharmacometrics/<http://www.farmb=
io.uu.se/research/researchgroups/pharmacometrics/>

From: owner-nmusers_at_globomaxnm.com [mailto:owner-nmusers_at_globomaxnm.com] On=
 Behalf Of Gewitz, Andrew
Sent: Friday, November 11, 2016 6:53 AM
To: Mark Sale
Cc: nmusers_at_globomaxnm.com
Subject: Re: [NMusers] $PRIOR with normal for OMEGA?

Mark,

Without getting too technical, recall that the inverse wishart is a distrib=
ution on matrices, and is naturally conjugate to the multivariate normal di=
stribution (i.e. on vectors of random variables.) Generally, this is chosen=
 because posterior computations are simplified and it is easy to sample fro=
m this posterior distribution.

What you're suggesting is the Matrix Normal distribution. Since it is not c=
onjugate to the multivariate normal, posterior computations can cause heada=
ches. So while it is *possible* to use such a distribution as a prior, it i=
s cumbersome to work with in practice and requires thinking about some thin=
gs like correlation and scale in non straightforward ways.




--------------
Andy Gewitz, PhD
Bioengineering and Therapeutic Sciences
University of California, San Francisco

On Nov 10, 2016, at 7:57 PM, Mark Sale <msale_at_nuventra.com<mailto:msale_at_nuv=
entra.com>> wrote:

Is it possible to use a normal prior for OMEGA? The default is inverse Wish=
art, but I'd be interested in using Normal (insuring that it is positive de=
finite) Any ideas?

thanks





Mark Sale M.D.
Vice President, Modeling and Simulation
Nuventra Pharma Sciences, Inc.
2525 Meridian Parkway, Suite 280
Durham, NC 27713
Phone (919)-973-0383
msale_at_nuventra.com<msale_at_kinetigen.com>
CONFIDENTIALITY NOTICE The information in this transmittal (including attac=
hments, if any) may be privileged and confidential and is intended only for=
 the recipient(s) listed above. Any review, use, disclosure, distribution o=
r copying of this transmittal, in any form, is prohibited except by or on b=
ehalf of the intended recipient(s). If you have received this transmittal i=
n error, please notify me immediately by reply email and destroy all copies=
 of the transmittal.



Received on Fri Nov 11 2016 - 09:24:16 EST

This archive was generated by hypermail 2.3.0 : Fri Sep 27 2019 - 16:52:36 EDT