Today is data collection day. I expect less than 50% survivorship of my samples after experimental manipulation on a good run. This is just due to the experiment itself, excluding other factors. On average, about 33% of the survivors are productive in terms of usable data. Which is why prep work is such a fucking beast - this is a numbers game and I routinely prep over a hundred samples, each one requiring an average of 4 minutes of attention. Of these, fewer than 50% turn out to be not-dead and correctly staged for the experiment, and these proceed for experimental treatment (from which there are survivors and non-survivors). So when I come in this morning and find that 10% of my post-experimental treatment samples succumbed to some kind of infection, thus further reducing the chances of getting good shit out of some non-zero number of samples, it puts me in a pretty foul mood.
ETA: OK, so I don't hate *everything*...it turns out that even amongst those tiniest percentages of survivors I got some fucking awesome outputs. w00tang!!!