HOME | PRESS RELEASES | HOW TO ORDER | CONTACT US | CAREERS | Japanese

Phoenix® NLME™ Features

Key features of Phoenix NLME include the following:

Model Development Tools

  • Leverage a library of built-in, configurable models to begin modeling quickly and easily.
  • Visualize and modify models graphically using the Drug Model Editor (DME).
  • Customize models textually using the Pharsight Modeling Language (PML).
  • Develop covariate scenarios from a base model using a powerful graphical user interface.
  • Compare model and scenario results both graphically and analytically.
  • Launch estimation engines from a command line, script, or batch file.

Native Graphics

Automatically generate a variety of figures for each run. Reproduce customized graphics by using workflow templates with each analysis. Figures include:

  • Latticed population and individual prediction with observations vs. time.
  • Goodness of Fit plots: CWRES, WRES, and PCWRES vs. independent, predicted, and time after dose.
  • Individual and overall predictive check.
  • Dependent variable vs. population and individual predictions, and time after dose.
  • Quantile-Quantile plots for random effects.
  • Random effect vs. covariate.
  • Random effect correlation plot.

Modeling Capabilities

  • Build models with explicit, ordinary differential equation, or custom likelihood function.
  • Model continuous and categorical data with continuous and categorical covariates.
  • Provide individual or group dosing schedule with repeated and/or steady state dosing.
  • Easily specify two levels of random effects through the GUI and more levels with PML.
  • Generate initial estimates with a pooled or two-stage run.
  • Utilize a variety of estimation engines, including: Naive Pooled for Gaussian and non-Gaussian responses; Iterated Two-Stage (IT2S-EM) for Gaussian and non-Gaussian responses; First Order (FO) for Gaussian responses; Extend least squares First Order with Condition Estimates and Interaction (ELS-FOCEI) for Gaussian responses; Lindstrom-Bates FOCE for Gaussian responses; Adaptive Gaussian Quadrature (AGC) for Gaussian and non-Gaussian responses; Laplacian for Gaussian and non- Gaussian responses; Quasi-random Parametric Expectation Maximization (QRPEM) for Gaussian and non-Gaussian responses, Non-parametric for Gaussian and non-Gaussian responses, with no assumptions regarding the random effects distribution.
  • Evaluate models with bootstrapping, likelihood profiling, and post predictive check.

Performance

  • Phoenix NLME uses the Argonne MPICH2, which takes advantage of significant speed gains available on modern dual- or quad-core computers.
  • Phoenix NLME is designed from the bottom up to support parallel execution on multiple processors within a single NLME run.
  • Dispatch jobs to a grid or cluster with optional connectors customized to your infrastructure.
  • NEW!  the new, easily configured Job Management Server frees up the user's computer by allowing the execution of Phoenix jobs (including entire workflows) remotely on Windows Servers.

Compatibility and Validation