Two software packages have been developed and distributed LYSIS and ADAPT. Both software packages are continuously expanded to incorporate new methodologies developed in Core Projects #1, #2 and #5. LYSIS is an interactive modular software package of programs performing high-level tasks that can be used for linear and nonlinear time-series analysis, and system modeling and simulation. The latest Version 6.2 includes the """"""""principal dynamic mode"""""""" analysis technique and a novel class of """"""""artificial neural networks"""""""", both developed in Core Project #1. These programs are unique to LYSIS and enhance significantly the data analysis capabilities in many biomedical applications. In addition, the unique LYSIS programs of Volterra kernel estimation were converted into MatLab toolset in order to promote their broader use among the peer community. During the past grant year we have continued the distribution and support Release 4 of ADAPT. Release 4 includes: extensive Model File Library containing important pharmacokinetic and pharmacodynamic models; complete error analysis for the maximum likelihood estimator in ID; information criteria for model selection with each estimator; expanded predictive statistics and residual analysis; user defined secondary parameter models; partially optimized sample schedules for adaptive trial designs; automated multiple subject parameter estimation. Release 4 is supported for the following hardware/software platforms: PC/Windows 95/NT (Digital Visual Fortran v. 5.0); PC/Windows 3.1 (DOS) (Microsoft PowerStation Fortran v 1.0); SUN/Solaris; VAX/OpenVMS; HP/HP-UX (available upon request). The User's Guide for Release 4 includes a tutorial introduction to modeling with ADAPT. Also, an extensive Model Library is available that includes over 35 pharmacokinetic and pharmacokinetic/dynamic models. These library Models are fully documented in the User's Guide and many are illustrated in the numerous example runs contained in the User's Guide.
Showing the most recent 10 out of 84 publications
