NONMEM Users Network Archive

Hosted by Cognigen

Re: RE: NONMEM 7 Update

From: A.J. Rossini <blindglobe>
Date: Mon, 6 Jul 2009 17:52:33 +0200

On Mon, Jul 6, 2009 at 2:35 PM, Alice Nichols<NICHOLA2
> The fact that NONMEM7 will not be open source code is very unfortunate,
>
> it has not been often,  but there have been occasions when by going =
over the
> source code I was able to understand aspects of NONMEM programming that w=
ere
> still unclear to me,
>
> being able to get a look at the actual algorithms in software had been a
> major advantage for NONMEM!!!!

I agree that having the source code available, even though released
under a restrictive license, can be valuable for a few people (myself
included). I would assume that the majority (at least 51%) of users
don't care, and probably shouldn't.

The comment about open source release of comparable tools -- well,
such do exist, and while some are suboptimal to NONMEM in their
current implementations, they could be improved if the right people
had time and interest -- but this is difficult, as the models and
statistical framework is fragile, given the pressure most of us (in
Pharma) have to use the minimal amount of information to get the most
robust answer feasible.

For example, the R packages nlmeODE, PKfit provide a starting point
which could be brought up to speed with the right people supporting
them, and other basic tools could be enhanced to provide a friendlier
environment.

People have pointed out Monolix, but that's more like NONMEM -- still
requires a commercial blackbox to run. No problem with blackboxes
for validation, SAS being the most famous.

Anyway, there are serious discrepancies between the published
algorithms and the implementations. So it isn't necessarily possible
to use the published work alone to figure out how to reconstruct
estimates except for straightforward model/data combinations.

unless the community is willing to put in (after-hours) effort for a
real open source solution, let's enjoy the strong support we currently
get from our commercial providers, and support and work with them.
After all, they are working to get our support (and funding!).

best,
-tony
(Stacey, for you:
 Group Head, M&S Statistics, Novartis Pharma AG).


blindglobe
Muttenz, Switzerland.
"Commit early,commit often, and commit in a repository from which we
can easily roll-back your mistakes" (AJR, 4Jan05).

Drink Coffee: Do stupid things faster with more energy!
Received on Mon Jul 06 2009 - 11:52:33 EDT

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.