Trail membrane ligand induces apoptosis to neighbor cells.
many individuals not responsive to trail therapy to kill cancer.
cell to cell variability.
caspace signaling part of kill signal
waiting delay varies dramatically. Distribution depends on dose.
time to die post signal is fixed.
survivors from 80% kill dose, regrow, dose again, same distribution.
prior variability? (Eg cell cycle position)
monitor time to death of sister cells. Very close correlation. Correlation decays exponentionally, hit background in 2 cell divisions.
Linear relation between gfp and antibody hard t find.
cov. 3 to. 4 log-normal distributed intensity data.
consequences of cell cell variability. Get dose response behavior from a binary process (reverse of Ferrie gradation from binary switch)
partial kill (in place of tumor stem cell)
Sensitivity depends on where you are in p space. Of course.
Look for divergence in model behavior for small differences in peace. Position in caspace 3 vs xiap separates two types of death.
Theme: phase diagrams.
Probablistic perspective.
Biochem grind and bind. Vs genetics that pipe’s important.
Bayesian. Infer marginal probability of parameter values given data. See if data effects distribution.
Treating parameters as independent distributions is fail. Duh. Best fit parameters so not lie at the individual peaks.
Should communicate the joint probability matrix of the data.
Role of data. Multiple methods drop uncertainty the most.
Models should be managed in rules not odes. Readable.