To evaluate a mutant effect, the null hypothesis is that the effect is the same with the control. In this case, I can partition the wildtype survival function into the bulk and limiting module with a mixing parameter. Both wildtype and mutant lifespan data can be fitting with the null model.
Alternatively, mutant survival functions can be fitting with different parameters.
Likelihood ratio test, AIC or BIC can be used. Alternatively, permutation of experimental data can be used to generate the null-distribution.
No comments:
Post a Comment