Parameter identifiability and redundancy in a general class of stochastic carcinogenesis models

Area of Science:

• Mathematical Biology
• Cancer Research
• Systems Biology

Background:

• Previous work established parameter identifiability for the two-mutation cancer model.
• Recent carcinogenesis models incorporate genomic instability, generalizing earlier models.
• This study extends identifiability analysis to these more complex, generalized cancer models.

Purpose of the Study:

• To investigate parameter identifiability in the generalized carcinogenesis models of Little and Wright.
• To generalize and extend previous findings on parameter identifiability to broader classes of cancer models.
• To identify specific combinations of identifiable parameters within these models.

Main Methods:

• Analysis of parameter identifiability in mathematical models of carcinogenesis.
• Generalization of existing identifiability results to new, more complex models.
• Numerical evaluations to assess the sharpness of derived bounds.

Main Results:

• Identifiability is shown for generalized cancer models, extending prior work.
• For a simpler model, at most two fewer parameters are identifiable than the total number of biological parameters.
• For a more general model, the number of identifiable parameter combinations is reduced by the number of destabilization types.

Conclusions:

• Parameter identifiability results are generalized to larger classes of quasi-biological cancer models.
• Specific combinations of identifiable parameters are identified.
• These findings have theoretical interest and practical significance for cancer model parameter estimation.