A repository for the python codes to carry out the ABC-SMC model selection and parameter inference in the paper "Competitive binding of STATs to receptor phospho-Tyr motifs accounts for altered cytokine responses", published in eLife (2021). The data files required for the analysis are also included along with the results, presented in the paper.
Specifically, there are three python codes in this repository which can be used to rerun the Bayesian statistical analyses in the paper. Firstly, the code ABC_SMC_model_selection.py can be used to replicate the result of the model selection, to determine between the two hypotheses relating to the way in which receptor molecules are internalised into the cell. The other two codes, ABC_SMC_RPE1.py and ABC_SMC_TH1.py perform the ABC-SMC parameter inference algorithm and can be used to generate the posterior distributions for the parameters in the mathematical models introduced in the materials and methods, section "Mathematical models", for the RPE1 cell data and the Th-1 cell data respectively. Details of the Bayesian model selection and parameter inference algorithms can be found in the materials and methods, section "Bayesian inference".
Also included in this repository are the normalised data files, using the normalisation defined in Equation (5) of the paper. These are the eight text files beginning IL6_ or IL27_. The results of the model selection are included for the 15 different distance measures between model and data, where "Accepted_models_iteration_14.txt" corresponds to the final iterations results, where the model and data are closest. Finally, the results of the ABC-SMC inference are also given as text files, "RPE1_posteriors.txt" and TH1_posteriors.txt", for the RPE1 and Th-1 cell data, respectively.
