michael_nielsen

@michael_nielsen however, strain variation and polymorphism complicate matters, along with so-called "pan genomes".

@michael_nielsen I personally think a good operational definition is "do sequences from these two organisms assemble together" :)

@michael_nielsen sort of a computational definition of sex. But I've never explored the idea.

@michael_nielsen if you write something on it, I'm happy to publicize it and join in with everyone in telling you why it's wrong. :)

@ctitusbrown on known solutions, then use the model to (quickly!) do future "simulations". Apparently useful for doing real time 3d graphics

@michael_nielsen you're assuming we have a good objective function. I'm not sure we do :)

@ctitusbrown So with my genome and a hippo genome the model would be trained to say "not the same species"

@michael_nielsen sure, but in addition to epistemological confusion, I would worry about sampling bias in training set.