ISoP ISoP on Twitter | ACoP ACoP on Twitter | PAGE | WCoP | PAGANZ | PAGJa

Use of Berkeley Madonna for Quantitative Systems Pharmacology (QSP) Modeling


Why this Tutorial ?

• This video tutorial, sponsored by the ISOP Quantitative Systems’s Pharmacology Special Interest Group, introduces the features of Berkeley Madonna for QSP via demonstration of this program with a published small scale QSP model

What is Berkley Madonna ?

• Developed by Robert Macey and George Oster of the UC Berkeley under the sponsorship of NSF and NIH
• Designed to numerically solve systems of ordinary differential equations (ODEs) as fast as possible
• Is arguably the fastest, most convenient general purpose differential equation solver available today
• Is a tool of choice for some researchers in engineering
• Is essentially an ODE solver with some add-on functionality
• Is not designed specifically for QSP modeling
• Is model building based on programming language, not graphic blocks

Why Berkley Madonna ?

• The diversity of QSP modeling purposes, scopes, and modelers’ preferences requires a diversity of QSP modeling tools
– Fit-for-purpose small scale models vs. comprehensive, large scale models
– Enhancing mechanistic understanding vs. predicting clinical responses to intervention of novel targets
– Focusing on mean behaviors vs. exploring population variability
– Graphical model building vs. programming languages

• Berkeley Madonna is a great tool for some QSP modeling needs
– especially for small scale models with a focus on mean behaviors

Model Used in Tutorial

• Developed by Lu et al. and published on Frontiers in Pharmacology
– Use of systems pharmacology modeling to elucidate the operating characteristics of SGLT1 and SGLT2 in renal glucose reabsorption in humans (
• Purpose
– To understand the contributions of SGLT1 and 2 to renal glucose reabsorption, thereby explain clinical pharmacodynamic observations of SGLT2 inhibitors
• Features of the model
– A fit for purpose, small scale QSP model, with emphasis on plasma glucose renal filtration, transfer along proximal tubules, and reabsorption via SGLT1/2, as well as competitive inhibition by SGLT2 inhibitors (dapagliflozin, canagliflozin)
– No component for glucose-insulin interactions, as clinical data were collected under glycemic clamping procedures
– Focused on mean behaviors, no consideration of population variability
– Includes healthy and type 2 diabetes subjects , but no construction of virtual patients