QSPC2020 Workshop canceled due to
All the best – don’t travel to much, wash your hands, and stay healthy!
The Workshop will be provided as a Free Webinar instead … shortened from 8 to 4 hours
Date: Tuesday, April 21st, 2020
Time: 2:30 pm – 6:30 pm Swiss Time
8:30 am – 12:30pm East Coast Time
>>> Click for Free Webinar Registration <<<
Advanced QSP Modeling in R
“From Clinical Data to Virtual Subjects, Cohorts, and Populations”
IntiQuan welcomes you to a full day hands-on workshop on advanced QSP modeling!
In the past two decades, quantitative systems pharmacology (QSP), a mechanistically oriented form of drug and disease modeling, has established itself as a powerful tool to quantitatively integrate data and knowledge. Its scope is to support the assessment of drug efficacy and safety problems in model-informed drug discovery and development. The impact of QSP in model-informed drug discovery and development continues to grow and is increasingly recognized within the pharmaceutical industry, from early stages in drug discovery to late-stage development and life-cycle management, up to support of regulatory submission.
Recently, an approach to a QSP workflow has been proposed, serving as a guide to data programmers and modeling scientists. This workflow covers the entire QSP data structuring and modeling process by providing a recipe with the ingredients needed for a QSP modeling activity to proceed.
The workshop is picking up on the workflow described in the paper and shows how to put the abstract concepts into application on a realistic QSP modeling example. Special emphasis is put on the estimation of individual parameters based on clinical data, allowing to determine Virtual Subjects, Cohorts, and Populations in an estimation-based approach that is suitable even for large scale models and a large number of patients in the considered data. The ability to estimate such individual level parameters in QSP models has the potential to open up advances in diagnostics through the use of advanced statistical methods, such as machine learning.
The webinar is split into three parts with the following topics:
Part 1: General model implementation and simulation
- Model representation and simulation:
- ODE based syntax
- Biochemical reaction-based syntax
- Standardized general data format applicable to both QSP and NLME modeling.
- Import of models from SimBiology, Berkeley Madonna and SBML into R.
Part 2: Realistic QSP modeling example
- Representation of example model and available clinical data.
- Sensitivity analysis aiming at identification of influential parameters in the model:
- General sensitivity analysis based on mean normalized sensitivity metric
- Targeted sensitivity analysis for specific model metrics of interest
- Stepwise Parameter Modeling (SPM) for determination of parameters for which information in the data is contained.
- Estimation of mean parameters on mean and individual level data.
- Stepwise Parameter Modeling – Inter-Individual Variability (SPM-IIV) for determination of parameters for which information about variability is contained in the data.
- Estimation of individual level parameters based on individual level clinical data
- Generation, representation, and simulation of:
- Virtual subjects
- Virtual cohorts
- Virtual populations
- Application of individual parameter estimates in Diagnostics (#Machine Learning, #Precision Medicine).
Part 3: Can I trust my model / the parameter estimates?
- Robust parameter estimation based on sensitivity equations and multi-start optimization in R.
- Analyzing models, informing modeling decisions, using profile likelihood and other methods.
- Which experiments should be conducted to be able to decide between different mechanistic hypotheses?
The webinar is designed as a hands-on tutorial. Each topic will first be presented on slides and will be illustrated based on realistic examples. Between topics, the participants will have the chance to implement and reproduce the different steps on their own using the provided example model.
After the webinar, the participants will have learned to write their own mechanistic QSP models, simulate models, and perform parameter estimation – with and without consideration of parameter variability. In addition, participants will have gained insight into how models can be analyzed to better inform the design of new experiments to increase mechanistic understanding and trust in the models.
Basic knowledge of writing scripts in R is an advantage but not strictly necessary.
Note that due to reduction from 8 (workshop) to 4 (webinar) hours there might be less time for individual hands-on experience. The webinar material is self explaining though and can easily be digested per the webinar.
Two weeks prior to the webinar date participants will receive detailed information for the setup of their computers or alternative access to the tools through web based means. All workshop material (slides, example files, etc.) will be made available as download.
Henning Schmidt, PhD, IntiQuan GmbH
Henning Schmidt is an expert in Model-Informed Drug Development with over fourteen years of industry experience. He has been supporting projects from target validation in the early phases of drug discovery to study design for post-marketing commitments. During the last decade, he has provided decision-making support to drug discovery and development teams in various therapeutic areas, including oncology, dermatology, immunology, respiratory, bone and muscle wasting diseases. This has led to several go/no-go decisions on the progression of novel drugs through their development cycle and successful registrations. In addition, Henning is active in the area of tool development for improving the efficiency, quality and compliance in the areas of systems biology, QSP, and pharmacometrics. Henning received undergraduate education at Darmstadt University, Germany and SUPELEC in Paris, France. He obtained his PhD in Control Theory and Systems Biology at the Royal Institute of Technology in Stockholm, Sweden. He has worked for Fraunhofer Chalmers Research Center (Gothenburg, Sweden) and Novartis Pharma AG (Basel, Switzerland) prior to founding IntiQuan in 2015.