Further refine dealing with dead cells

[aarmey]

We have more death observations now that we're not throwing out cells. We should evaluate how they're distributed. It looks like the times are potentially exponentially distributed, which would make fitting them much easier.

I'm not sure we need to include this in the current paper, but the death times could be interesting as follow-up on their own.

[Farnazmdi]

Forgot to mention it here, I fitted an exponential to each concentration in #676 in the figure13. I can try and write a distribution function for them and add them to the likelihood. I remember you said it is not urgent for now, please let me know if it is a reasonable time to start working on it, or there are other priorities.

[aarmey]

I think it's reasonable to try. It could be good to get Readthedocs working... I don't think there's one urgent thing.

[Farnazmdi]

note for me to remember: the last time we visited this, we figured that since the distribution of death times is not independent of the phase duration times, due to censoring, it is not very straight forward to pursue this.