QualitativeModelFitting

QualitativeModelFitting (qmf) is a package designed for validating a model against arbitrary observations. The concept stems from that of unit testing in software development. Using qmf, each part of a model is tested by statements derived from literature or in house data. These statements are encoded as a qmf input string which is used together with an antimony string as input to the qualitative_model_fitting.Runner class.

Click below for more information on usage.

• The qmf input string
• Runner

This is the first version of qmf and there are a number of planned features that are not yet supported. In no particular order, these are:

Todo

• Build in full profile type analysis using a machine learning classification model. This would allow for profiles to be compared agaist (e.g.) a hyperbolic, transient or sigmoidal curve.
• Implement a cache system for performance improvements
• Implement the ‘between’ operator for implementing a rule that a component should be between x and y.
• Implement the ‘almost’ operator for floating point comparisons
• Implement the ‘start’ and ‘end’ operators for time intervals to abstract the need to always remember the end point of a simulation
• Allow for assigning variables to collections so we can list species that have the same rules
• Build in loops so we can do bulk validations
• Build the steady state block
• Build a dose response block
• Build the sensitivity block
• Build a plot block