website·Indexed 31 days ago
https://nmguides.vrognas.comTechnical reference and user guides for the NONMEM system, covering versions 7.2 through 7.6.0. Includes detailed documentation on NM-TRAN control records, estimation methods (FO, EM, MCMC Bayesian), differential equation solvers like LSODA (ADVAN13), probability density functions, and utility programs for analyzing $COV SIR data. Provides guidance on clinical trial design, nested random levels, and matrix algebra tools for parameter estimation.
OMEGA.NM-TRAN (NONMEM Translator) is a FORTRAN 90/95 preprocessor used before a NONMEM run. It simplifies the user's workflow by translating user-friendly inputs into the specific formats required by NONMEM. It produces three primary outputs:
PRED) that must be included in a NONMEM load module.While NONMEM can be used without NM-TRAN, the preprocessor is strongly recommended, especially for beginners, as it handles the rigid formatting and complex subroutine generation automatically.
The nmhelp system provides technical documentation for NONMEM 7.6.0, organized by file extension. Users can access these keywords via the command line using nmhelp <keyword>. The documentation is categorized as follows:
*.ctl): Documentation for records such as $THETA, $OMEGA, and $DATA.*.dat): Reference for data-record fields like AMT, EVID, MDV, and TIME.*.exa): Reference example control streams.*.fil): Conventions for NONMEM input and output files.*.for): Runtime variables including ICALL, NEWIND, and COMACT.*.gen): High-level explanations of individuals and error models.*.out): Details on items found in the NONMEM output/report file.*.ppp): Documentation for ADVAN, SS, dose options, and TRANS.*.stb): Stubs for CCONTR, CONTR, CRIT, MIX, and PRIOR.*.sup): NM7+ features including parallelism, XML, and table tools.*.utl): Reference for functions like RANDOM, PHI, CHOL, GAMLN, and SIMETA.NONMEM automatically creates several files based on the root name of the control stream file. The delimiter and numerical format of these files are determined by the $EST … FORMAT= specification. Key output files include:
root.cov: Full variance-covariance error matrix for THETAs, SIGMAs, and OMEGAs.root.clt (NM74): Lower triangular form of the variance-covariance error matrix in TOSL order, intended for use as prior information in control stream files.root.cor: Full correlation matrix for THETAs, SIGMAs, and OMEGAs.root.coi: Full inverse covariance matrix (Fisher information matrix) for THETAs, SIGMAs, and OMEGAs.root.grd (NM72): Gradient values for classical NONMEM methods.root.xml (NM72): XML markup version of the NONMEM report file.NONMEM is a batch-oriented computer program written in FORTRAN 90. It consists of several key components:
PRED subroutine in NONMEM, specifically designed for population pharmacokinetic (PK) data analysis. While its formal name in the calling program is PRED, it is informally known as PREDPP. It can be used for population analysis, single-subject analysis, and simulation of PK data.Abbreviated code is used in NM-TRAN control records (e.g., $PRED blocks) to specify the computation of model quantities. It is syntactically similar to FORTRAN, primarily using assignment statements, conditional statements, and arithmetic/logical expressions.
Key characteristics include:
;).PRED to compute predicted values and partial derivatives.PRED subroutine in NONMEM, specifically designed for population pharmacokinetic (PK) data analysis. While formally named PRED (the entry name used by NONMEM calling programs), it is informally called PREDPP. It is used to compute model-based predictions for observations and can be applied to population data, single-subject data, or simulations.OMEGA.NONMEM (Nonlinear Mixed Effects Model) is a software package written in FORTRAN 90/95 used to fit general statistical nonlinear regression-type models to data. It is specifically designed for population pharmacokinetic (PK) data analysis, allowing for the modeling of both fixed effects (measured concomitant effects) and random effects (unexplainable inter- and intra-subject effects).
The system consists of three primary components:
PRED) to obtain predicted values and partial derivatives.PRED for Population Pharmacokinetics): A specialized PRED subroutine distributed with NONMEM. It contains a library of pharmacokinetic models and dosing regimens, removing the need for users to write their own FORTRAN subroutines for standard PK analysis.$) into the specific input files required by the NONMEM engine.