2 replies to this topic

[olivier levionnois]

Hello all,

 

Do you know if it is possible to run population modeling on a user-model written in WNL ASCII Code ? I can't find the way.

 

Many thanks !

Olivier

[Ana Henry]

Hello Oliver,

 

The short answer to your question is NO.

 

The long answer:

Population PK/PD is performed with our product Phoenix NLME. You only need a valid license to run NLME if you have already installed Phoenix WinNonlin. Without NLME you cannot do non-linear mixed effect modeling.

 

Even with Phoenix NLME you would need to translate your WNL Ascii user model into PML (Pharsight Modeling Language) code. Take a look at some examples to get familiar with NLME. My recommendation is to create a Phoenix Modeling object then select the option 'Set WNL model', this list all the same library model that you had available in WNL classic. Make sure that the Population option is selected. Please note that these models will have secondary parameters pre-specified and depending on your model modification this might or not work. Since you probably wrote a user-model for a reason that library model would need modifications so then select 'Edit Graphical' and try to include your changes graphically. If that is not possible then select 'Edit textual' and change the code. The code is not the same as WNL ASCII so you will need to learn a bit about PML. However, once you get used to it you will have a lot more flexibility.

 

 

Ana Henry

[Simon Davis]

Small update, since Phoenix 6.4 you can run a Naive Pooled population run in Phoenix with the standard WinNonlin license, i.e. the NLME license is not required for this simpler option;

 

Naïve pooled engine Applicable to Gaussian and user-defined log likelihood data. The Naïve pooled engine, when applied to population data, treats all observations as if they came from a single individual in that it ignores inter-individual variations in ETA values. All ETAs are forced to zero, and no OMEGA parameters are computed - only THETA and SIGMA. The engine can also be applied to a single individual, or to individuals separately as a series of individual fits in a multiple individual dataset, in addition to application to all individuals collectively in a population model. When applied to all individuals in population mode, the engine pools the data for evaluation into a single individual log likelihood function that contains no random effect parameters, but respects inter-individual differences in dosing and covariate values.

 

The engine minimizes the exact negative log likelihood, either as a Gaussian or user specified function. No approximations are necessary since there is no population distribution and hence no joint likelihood to integrate. The minimization is performed by the same quasi-Newton algorithm as used in the other engines. As with FO, FOCE-ELS, and Laplacian, iterations simply correspond to iterations of the quasi-Newton optimization algorithm. Also as with FO, in principle only a single pass through the quasi-Newton method is required, but the PHX implementation repeats the optimization from the final value of the previous run until successive runs yield the same log likelihood to within a tolerance of 0.001.

Edited by Simon Davis, 22 March 2017 - 01:36 PM.