News for hackers. Stuff that matters.
Re:Chemical space is really big (2014) by whatshisface on Friday June 25, 2021 @06:23PM (#27637003)
Molecules obey known laws of physics and can in principle be simulated exactly. That is not practical with present-day computers because it's quantum-mechanical and has an exponentially large state space. Heuristic and approximate methods are used to pare this down, sacrificing absolute truth, leading to results that are not very reliable. That is why experiments are still done in chemistry labs even though everything that happens in a chemistry lab has been "understood" since the 1930s.

Chemists focus in on the least simulatable problems because most interesting chemistry happens right on the border of not happening at all. Molecules that are very easy to calculate are ones that small energy errors don't matter for. That makes them either incredibly stable or incredibly unstable, but chemistry happens near the boundary.

Re:Chemical space is really big (2014) by dmurray on Friday June 25, 2021 @08:40PM (#27637953)
I upvoted this because it's an interesting take, but it doesn't feel like it's true. Chemists pick new molecules to experiment with primarily because they're similar to old ones that solve a particular problem (cure a disease, colour an emulsion, explode very fast...), not because they tried computational simulation of those molecules and couldn't overcome precision/chaos issues.
Re:Chemical space is really big (2014) by whatshisface on Friday June 25, 2021 @09:33PM (#27638225)
If a molecule solves a particular problem, then it is neither so unstable that a poor simulation could predict its instability, nor so stable that a poor simulation could predict its stability. The molecules that are the easiest to predict computationally tend to be the least interesting. The energy difference between the bound and unbound states has to be small for it to be interesting which automatically puts it in the realm of difficult simulation. So no, chemists do not pick difficult molecules on purpose - they do so because difficulty and interestingness originate from the same place.
Re:Chemical space is really big (2014) by 41b696ef1113 on Friday June 25, 2021 @09:01PM (#27638059)
>...not because they tried computational simulation of those molecules and couldn't overcome precision/chaos issues.

Existing simulations are slow and inaccurate. Why waste your time generating computational results that you do not trust when you can just iterate on existing molecules and have them in hand, ready for testing?