Introduction to modeling

The premise to all modeling is that, based on experimental observations, we believe that a set of rules gobvern the behavior of a system. Modeling the behavior of such a system involves the elucidation of the rules that govern the system to understand our observations and the use of such rules to further predict the behavior of a system under a range of conditions. Thus, models can be used to both explain or predict the behavior of a system given a set of conditions. The best models can perform both tasks satisfactorily. A simple example is the colloquial story of Newton’s apple. The observation was that the apple fell on Newton’s head. He derived the simple yet incredibly powerful whereby he observed that the Force, , applied to an object of mass, , resulted in an acceleration, . We now know that this model holds for most conditions in every-day activities but we know that it fails for e.g. relativistic effects. Therefore a model has a domain of application and a limited usefulness. However, a successful model can be employed accurately for both the explanation and the prediction of a system. In the case of cell-molecular biology, we aim to develop models that describe the behavior of cellular systems. The model can guide us to understand what are our gaps in the observations that prevent us from generalizing a theory and, when they capture the key significant aspects of the behavior of a system, predict the outcome of the behavrio of a system under a given set of conditions.

PySB as modeling tool

PySB is a set of software tools that enables users to develop, implement, and execute biological models in the Python programming environment. One of the main advantages of PySB is that it leverages the power of a very powerful programming language to express biological concepts as parts of a program. The properties of the programming environment are therefore the same properties found in PySB. Python is an object-oriented programming language that provides a useful environment for programming techniques such as data abstraction, encapsulation, modularity, message-passing, polymorphism, and inheritance to name a few. In addition to these technical advantages, we chose Python due to its readable and clear syntax. In our view, one of the most difficult issues with current biological modeling is shareability and transparency, both of which are addresssed, at least in part, by a clear syntax and a programmatic flow of ideas. PySB can handle simple models, modular models, and multiple instances of models, as shown in the tutorial. We invite users to contribute and share their innovations and ideas to make PySB a better open-source tool for the programming community.

A quick example

Using and running PySB can be as simple as typing the following commands in your Python shell. Go to the directory containing the file simplemodel.py (usually pysb/examples) and try this at your shell!:

[host] > python earm_figures.py

You will see some feedback from the machine, depending on your operating system (and assuming PySB is correctly installed). After a few seconds of calculations you should get two figures. The first figure shows the experimental death time determined form experiments (as dots with error bars) followed by the model-predicted average (solid line) and the standard deviation ranges (dashed lines). The second graph will show you the model signatures of three species, namely initiator caspase (IC) substrate, effector caspase (EC) substrate, and mitochondrial outer membrane permeabilization (MOMP) as indicated by release of Smac to the cytosol. You have now run a model! Feel free to open the files earm_1_0.py to see a simple model instantiation and earm_figures.py to see how the model is run and the figures are generated. If you want to learn how to build biological models in a systematic (and we think fun) way, visit our Tutorial.