Rule-based modelling

Teachers: Vincent Danos (, Jérôme Feret (, Jean Krivine (

Schedule: 8 Sessions of 3 hours each (2 hours of lectures, 1 hour of practical works)

Evaluation: Three small take home assignments on three subparts of the module. One question on this course may be included in the big take home exam of CompBio I, Synthetic Biology and Systems Biology.


Overview of the course:


I) Rule based Modelling-part

In this class, you will learn a new way, called rule-based modelling, to perform mechanistic modelling which is very different from that of the more usual approach through ODEs.

This approach builds models by writing, in a simple formal language called Kappa, rules that describe directly the possible interactions between proteins; these models can be equipped with rate laws and then simulated and analyzed with open source software tools (that we will provide). The class is primarily practical where you will develop, think about and analyze a series of increasingly complicated systems using Kappa. There is also some pure lecturing time in order to explain the novel analysis tools offered by Kappa.


Course objectives (I):

The main aim of the class is to help you to develop your intuitions about how signalling networks operate. The use of rule-based modelling makes it easy to try out variants of a model, which sometimes have very different behaviour, and so there is a strongly emphasis on exploratory  learning. You will also learn some specific technical material, notably about stochastic simulation, that complements what you have learned about ODE-based deterministic simulation in other classes.


Requirements (I):

There are no absolute prerequisites for this class although a passing familiarity with programming might ease the initial learning curve of writing rules and executing models and some basic intuitions about probability theory will help for understanding how simulation works.


Suggested readings (I):

– Vincent Danos, Jérôme Feret, Walter Fontana, Russell Harmer,& Jean Krivine. Rule-based modelling of cellular signalling. Invited in International Conference on Concurrency Theory (CONCUR 2007), number 4703 in Lecture Notes in Computer Science. 2007, © Springer.


II) Whole Cell Modelling part

In this course students will learn about and implement a physiological model of a cell then propose their own improvements and additions inspired by their research interests.

A large amount of work has been devoted to the mathematical and computational modelling of specific cellular processes. As accurate as these models may be, their isolation from the physiological cellular context hampers the study of the role they can play in global cellular behaviours. A whole ­cell model is an aggregate of mathematical representations of cellular subprocesses (e.g. translation, protein maturation, etc.). Of course, such submodels need to be validated against experimental data. Eventually, we expect the aggregate model to explain high level behaviours of a cell like the growth rate. During this hands-on workshop, such a model will be realized.

Format: Participants will be divided into several groups. In part 1 of the course, the groups will implement a base model of a cell then they will propose extensions to the model based on their research interests and expertise. This will be the occasion to ask questions, draw and test hypotheses about very exciting emergent phenomena at both the cellular and population scales, in a research-­like fashion. In part 2 of the course, one week later, students will complete their implementations and present their extensions to each other.

The models will be written and simulated in MATLAB.