PyCoTools

Package was developed by ciaran welsh (c.welsh2@newcastle.ac.uk) from Newcastle University, UK.

Installation

Use: pip install PyCoTools

Version Notice

This package was developed on Python 2.7 and Copasi version 16. Since version 16 three Copasi versions were released in quick sucession and some featuers about how copasi reads the CopasiML have apparantly changed. For this reason the current release of PyCoTools works on Copasi Version 16 and previous releases. Efforts will be made to develop a PyCoTools for the newest version of Copasi (version 19 at the date of writing this)

Tutorial

A full tutorial on how to use PyCoTools can be found here

pycopi

The pycopi module is a set of python classes for interfacing with copasi via python.

This can be used when a user wantes to setup an anaysis not native to COPASI, as exemplified by the pydentify2 module.

Features include:
    - GetModelQuantites     - Class to retrieving information about model 
                              and model entities
    
    - Reports               - Automate the setting up of specific and commonly used reports
    
    - ParameterEstimation   - Automate the setting up of a parameter estimation in copasi. 
                            
                            - Run a parameter estiamtion through copasi Via python and visualize 
                              the results
    
    - ExperimentMapper      - Automate the process of mapping model species to experimental data. 
                              Used by ParameterEstimation class
    
    - TimeCourse            - Simulate a time course using Copasi. Parse this data into python 
                              and plot it optionally. 
    
    - PhaseSpace            - Produce plots in phase space.
    
    - Scan                  - Automate the process of setting up a parameter scan. All three 
                              scan, repeat and random number generation) are supported. 
    
    - Run                   - Submit copasi file to CopasiSE. Unticks all 'executable boxes' then ticks 
                              the executable box of interest. 
   
    - InsertParameters      - Take parameters from pandas.DataFrame, python dict or from csv/txt files (output from copasi) 
                              put them into the model. Note: A known bug exists where if you have parameter 
                              sets in your copasi file then these get changed instead of the active parameters. 

PEAnalysis

The PEAnalysis module provides a number of convenient classes for visualizing and analysing parameter estimation data. The module can be used with any PE data regardless of origin but has been developed with copasi in mind.

Features Include:
    
    - ParsePEData           - Read data from a single parameter estimation file (Copasi output) 
                              or a folder of parameter estimation files
    
    - WritePEData           - Output Parameter estiation data back to file in more convenient format
    
    - TruncateData          - Parameter estimations often find local minima. These can be truncated 
                              from further analysis by either specifying how many percent (i.e. top 1%)
                              or what value you would like to use as cut-off, i.e <1e2). 
                              This class is used implicitely within the plotting classes and an 
                              interface is provided as an option.
    
    - PlotHistogram         - Plot histogram for each parameter in your estiation file
    
    - PlotScatters          - Plot all combinations of scatter graph
    
    - PlotBoxPlots          - visualize variance with box plots
    
    - PlotOptimizationPerformance - evaluate optimization performance with plot of (ordered) 
                                    likelihood Vs iteration. The flatter the graph the better. 

Pydentify2

The pydentify2 module uses the pycopi module to fully automate the process outlined by Schaber 2012: Easy Identifiability Analysis in Copasi.

Features Include:
    
    - ProfileLikelihood     - Setup and run profile likleihood for a model around current parameters
                            
                            - Optionally take parameters from a file (copasi output file for example) 
                            and perform profile likelihood around these parameters
                            
                            - Optionally take multiple sets of parameters from parameter estimation output 
                            file and perform profile likelihoods around each of them. 
                            
                            - Optionally run on a cluster (sun grid engine is implemented but modular 
                            nature of the code means this bit can be replaced by code for any job scheduler)

    - Plot                  - Plotting facilities for profile likleihood output. 
    
                            - Chi2 based confidence interval is automatically calculated and 
                            plotted as red dotted line for determining identifiability status. 
                                
                                -   This calculation requires the number of data points used in the parameter estimation, 
                                    the number of parameters estimated and the residual sum of squares 
                                    value for the current point in parameter space and your data. For this reason 
                                    when using the Plot feature with a profile likelihood calculated around parameters 
                                    already in the model, these values need to be suplied. When parameter estiamtion 
                                    data is used from file, the Plot class determines them for itself.