I studied bioinformatics and found the standard textbook, Albert's "Molecular Biology of the Cell"[0] to be one of the most captivating books I've read. It's like those extremely detailed owners' manuals for early computers, except for cells.
The amount of complexity is just absolutely insane. My favourite example: DNA is read in triplets. So, for example, "CAG" adds one Glutamine to the protein it's building[1].
There are bacteria that have optimised their DNA in such a way that you can start at a one-letter offset, and it encodes a second, completely different, but still functional protein.
I found the single cell to be the most interesting subject. But of course it's a wild ride from top to bottom. The distance from brain to leg is too long, for example, to accurately control motion from "central command". That's why you have rhythm generators in your spine that are modulated from up high (and also by feedback).
Every human sensory organ activates logarithmically: Your eye works with sunlight (half a billion photons/sec) but can detect a single photon. If you manage to build a light sensor with those specs, you'll get a Nobel Prize and probably half of Apple...
"The distance from brain to leg is too long, for example, to accurately control motion from "central command"
As a dancer, I have been fascinated by that fact. It means that dancers do not dance to the beat as they hear it - it takes too much time for the sound to be transformed by the ear/brain into an electrical pulse that reaches your leg. Instead, all dancers have a mental model of the music they dance to that is learnt by practice/repetition.
Dancing is just syncronizing that mental model to the actual rhythm that is heard. When I explained that to a bellydancer friend she finally understood the switch that she had made from being a beginning dancer to an experienced dancer who 'dances in their head'
You can clap your hands to a calibrated delay from the previous beat that you heard (predicting the next beat before you hear it). This is analogous to the principle of a phase-locked loop, which gradually adjusts an internal oscillator until it matches an external frequency. That internal oscillator can emit a beat just before the real one, offset just enough to cancel all the delays in the processing path.
This only works if the beat you're hearing is sufficiently stable.
Yeah, you often send commands several beats in advance. And then there's some lag too, because muscles are fairly viscous and take a bit of time to start up. You're basically dancing in the future, because you are behind. I think we just run pre-baked programs (from a lot of practice) and adjust their timings on the fly every few beats or a bar.
The Albert's MBoC is pretty much known as the reference textbook where I studied.
Note that the 4th edition is (sortof) freely available at the NIH website. The way to navigate through that book is bizarre though, as the only way to access its content is by searching.
The amount of complexity is just absolutely insane. My favourite example: DNA is read in triplets. So, for example, "CAG" adds one Glutamine to the protein it's building[1].
There are bacteria that have optimised their DNA in such a way that you can start at a one-letter offset, and it encodes a second, completely different, but still functional protein.
I found the single cell to be the most interesting subject. But of course it's a wild ride from top to bottom. The distance from brain to leg is too long, for example, to accurately control motion from "central command". That's why you have rhythm generators in your spine that are modulated from up high (and also by feedback).
Every human sensory organ activates logarithmically: Your eye works with sunlight (half a billion photons/sec) but can detect a single photon. If you manage to build a light sensor with those specs, you'll get a Nobel Prize and probably half of Apple...
[0]: https://amzn.to/2zzDt8P
[1]: https://en.wikipedia.org/wiki/DNA_codon_table