SAAM II

1. Pharmacokinetics and Pharmacodynamics (PK/PD) Research:

• Optimization of drug dosing regimens for enhanced therapeutic outcomes.
• Modeling drug absorption, distribution, metabolism, and excretion in the body.
• Studying drug-drug interactions and predicting their effects.
• Indirect and custom PD.

2. Population Pharmacokinetics:

• Analyzing drug responses across diverse patient populations with Two-stage methods.
• Personalized medicine: Tailoring drug dosing based on individual patient characteristics.

3. Systems Biology:

• Modeling complex biological networks and cellular processes.
• Understanding signal transduction pathways and regulatory networks.
• Investigating disease pathways and molecular interactions.

4. Biotechnology:

• Design and optimization of bioprocesses for pharmaceutical and biotechnological industries.
• Predicting the behavior of bioreactors and biocatalysts.

5. Metabolic Diseases Research:

• Studying metabolic disorders and their underlying mechanisms.
• Analyzing glucose-insulin dynamics and its relevance in diabetes research.

6. Tracer Studies:

• Quantitative assessment of the kinetics of radiolabeled compounds.
• Investigating tracer distribution and clearance in the body.

7. Experimental Design:

• Designing optimal experiments to gather data for parameter estimation and model validation.
• Assessing the sensitivity of model parameters to different experimental conditions.

8. Biological Modeling in Education:

• SAAM II as an educational tool for teaching bioscience and systems modeling.
• Demonstrating concepts of pharmacokinetics and systems biology in academic settings.

9. Peer-Reviewed Publications:

• SAAM II is used in various research studies and is cited in more than 50 peer-reviewed scientific journals per year (Google Scholar).

Notably, the glucose-insulin Minimal Models that are used in clinical trials to quantify insulin improvements of antidiabetic treatments, are implemented in SAAM II.[2]

In the early 1950s, Mones Berman and others at the NIH worked on problems in radiation dosimetry. Mones decided that compartmental models (systems of differential equations) were the best way to analyze the transient (kinetic) data being collected. He started the development of a software tool that eventually became known as SAAM. The power of SAAM was its dictionary that made it possible for a user to sketch their model, and then using the dictionary and a set of rules, create an input file directly from the sketch. SAAM took this information and created the system of differential equations that described the model. This meant that the user could think about biology/pharmacology while the program did the mathematics and statistics behind the scenes. It was a very popular program, but one had to visit the NIH and work with Mones to learn how to use the program.

Between 1986 and 1994, the University of Washington working through its Resource Center for Kinetic Analysis in the Center for Bioengineering, led by Prof David Foster with the help of Loren Zech from NIH, rewrote code including a strategic user interface, which led to SAAM II. The first version was released on the SUN in 1993. The PC version was released in 1994. Through several grants, in the 2000-2012 period, Foster and Vicini worked on generating the modern version 2.1, including a population analysis add-on called popKinetics. In 2012, the Epsilon Group, a Medical Automation Company in Virginia licensed the commercial rights to improve and distribute the software.

In 2022, the commercial rights to develop and distribute SAAM II software up to current version 2.3.3 were licensed to Nanomath LLC, a consulting and software company headquartered in Washington. The leadership and management of SAAM II were assumed by Simone Perazzolo, a scientist with experience in computational modeling of biological and pharmacological systems.

SAAM II utilizes three types of integrators for Ordinary Differential Equation (ODE) solving:

• RK 4-5th order: A 4th-5th order Runge-Kutta methods, which is a numerical technique for solving ODEs.

• Pade approximation of the matrix exponential: This method is based on approximating the matrix exponential using Pade approximants, providing an efficient approach for solving ODEs.

• Rosenbrock methods: SAAM II employs typical Rosenbrock methods, which are a class of implicit Runge-Kutta methods suitable for stiff ODEs (those with widely varying time scales).

SAAM II employs parameter optimization for multiple data fitting, utilizing a modified nonlinear weighted non-linear least-squares method derived from the Gauss-Newton algorithm. In regression tasks, users have the flexibility to create a weighting scheme based on either the error in the data or the model.

Additionally, SAAM II offers a Bayesian Maximum A Posteriori (MAP) option, allowing users to explore Bayesian parameter estimation. This feature enhances the analysis by incorporating prior knowledge and uncertainty into the parameter estimation process.

To assess the reliability of parameter estimates, SAAM II provides posterior and practical identifiability features. These utilize Fisher's information matrix and covariance matrix of the estimates to evaluate the quality of parameter identification, also in case of complex structures and numerous unknown variables.

Furthermore, SAAM II includes local parameter sensitivity, batch analysis, and in silico populations features, both of which are convenient tools for gaining insights into the model's behavior and assessing the impact of parameter changes on model outcomes.[3]

SAAM II can be found in curricula in many American and worldwide institutions, such as engineering, physics, and pharmacy schools.