I was wondering how a molecule transforms from one conformation into another. Surely, since the conformations are separated by a potential barrier, some activation energy must be required. The energies related to a molecule are translational, rotational, vibrational, and electronic. Out of these, translational and rotational motion do not change the intrinsic coordinates of a molecule, and these won't bring about a conformational change. The electronic transitions are too fast on the conformational time scale. This leaves, at least for an independent molecule, vibration as the only candidate for bringing about conformational transformations. Although collisions may also be considered, my focus is on independent molecules. I tried searching for a relation between conformations and vibrations but couldn't find any relevant results.
-
$\begingroup$ If you are talking about the conformational changes like the one that happens when a carbohydrate molecule like glucose goes from 4C1 to 1C4, I think it is mostly due to the rotations along the bonds in the molecule or at least that is how they are modeled. You can check this paper for further details. $\endgroup$– Hemanth HaridasCommented Oct 30, 2023 at 12:04
-
$\begingroup$ If you run a potential energy surface calculation, you will get that the system have minima representing each conformer, so, in other to got from one to another, you need to add energy in order to "break" the energetic barriers. As you will always have vibrations, they cannot make the system to undergo conformation changes. $\endgroup$– Camps ♦Commented Oct 30, 2023 at 12:27
-
2$\begingroup$ You can't really divide rotational vs vibrational movement for a floppy molecule $\endgroup$– Susi LehtolaCommented Oct 30, 2023 at 17:26
-
$\begingroup$ @KarstenTheis I have included the link to cross posted question on CSE. $\endgroup$– anantaCommented Nov 3, 2023 at 15:51
-
1$\begingroup$ Look at crest: github.com/crest-lab/crest It is ready to use, fast, and published. They also explain the background there. $\endgroup$– Martin - マーチンCommented Nov 5, 2023 at 10:32
1 Answer
Yes, you can think of conformations as arising from vibrational modes, typically low-energy ones. In fact, some conformational search methods use low-energy dynamics or projection to find new conformers.
I think Paul Hawkins still has the best review: "Conformation Generation: The State of the Art" (2017)
Stochastic methods based on Monte Carlo-simulated annealing (23) (MC) are often faster than MD methods. (24, 25) Methods based on low-mode or normal-mode searching attempt to retain the sound physics of MD while requiring still less computation time. (26-29) However, all these methods can still require many CPU minutes to produce conformations for a single molecule and are therefore unsuitable for processing even a few thousand molecules. Additionally, for MD and MC approaches (but, to a lesser degree, low-mode MD methods), the input conformation, which is required in all these methods, can bias the sampling, requiring the use of either multiple input conformations or longer sampling time.
Thus, many conformer search techniques either drive torsional angles or use distance-geometry (e.g. RDKit / ETKDG) as ways to stochastically project new conformer minima without needing MD or biasing the search.
