Notes on Transcription modeling:
need to be careful with discrete nature of Poisson cdf — incomplete gamma functions are not discrete, can’t throw away the floor function. So you do need to wait at least until you have 1 bcd molecule, probability of seeing less than x events in a Poisson process with rate (i.e. mean) v is gammainc(floor(x),v), where the floor is important. Probability of seeing less than 1 event is ~40%. Code for cdf of Poisson still needs work — not returning correct values.
Discuss with Thomas bcd sensing problem / hb transcriptional integration. Needs further discussion, wait until Princeton visit?