how would your represent a protein programmatically?

  • 1 Replies
  • 919 Views
how would your represent a protein programmatically?
« on: May 24, 2023, 12:23:08 pm »
Now anyone of use could duck-duck-go different answers on how bioinformatic groups are creating data structures/patterns/etc to represent proteins both in databases and in terms of how proteins interact with each other.

But I'd like to see what different people come up with using as little research as possible.  If you already know one or more methods from previous research, feel free to share, but maybe give other people a chance to comment first.  Most of us will have to do a little research, but once you have an idea, try to figure out the rest yourself.  I'll start by giving the simplest answer that I can think of: a linked list of enums which hold one of the values A, G, C, and T with a set of methods that can compute how different values can fit together.  (I think this is the approach that one of google's competitors took, I have to look it up though).
aka luke

Re: how would your represent a protein programmatically?
« Reply #1 on: May 28, 2023, 03:47:04 pm »
(you said as little research as possible, so I'm intentionally not checking my work.)

ACGT are the codons for DNA tho, not proteins? These four may make up some of the proteins, but I'd guess they're only four of many, and maybe we only use those four for enzyme construction instructions.  I know we (animals and fungi) digest other protein from other creatures by denaturing the proteins used in their construction, and we make more DNA from those proteins, but surely there are other proteins that can't be use for making DNA/RNA.  Like, we also consume starches and lipids, surely we take those apart and recombine them into new molecules.

I also remember from the "folding @ home" project that it's not enough to know the sequence of proteins, we also need to know their shape in space.  Hell, I remember that proteins are asymmetrical, and the property of "chirality" means that a mirror-image of a protein molecule could be biochemically inert to us -- "left-handed" proteins are useless to us if digested.  So a linked-list can't be enough.