Amylose parameters

[JakaBrklanje]

Hi all,

I've been using polyply with Martini 3 parameters to create some simulation systems of variable length line polymer solutions. I'm happy to say that CEL and DEX parameters worked very nicely for me so far. Now I wonder whether there are also amylose parameters available to be used with your software. Due to similarity to cellulose, I suppose I could use the same beads and only alter the angles and dihedrals to switch 1,4 beta to alpha orientation. However, I'm not sure on how one could validate the parameters obtained that way. This is why I've decided to ask here first.

With Regards, Jaka

[fgrunewald]

Dear @JakaBrklanje,

It's nice to hear that the Cellulose and Dextran parameters work well. To answer your question there are no amylose parameters available at the moment. But you are correct that the bead mapping and assignment would be the same as for Cellulose. Switching from beta to alpha connectivity will however change bond, angles, and dihedrals. To obtain parameters you would have to run an all-atom reference simulation, map it to CG resolution, and then fit the bonded parameters. That process is in principle not terribly difficult. However, amylose has the particular challenge that it forms helical structures stabilised by specific hydrogen bond interactions. Thus it might be necessary to apply a bias to the structure in order to keep it helical - like it is done for proteins.

If you're interested in obtaining parameters yourself check out the different discussion topics we have - especially this one and this one. We can also give advise on this particular sugar polymer when you're opening a new discussion topic.

Cheers, Fabian

[JakaBrklanje]

Dear Fabian, thank you for the explanation!

My main interest is to explain certain polymer properties of common and well-explained oligo- and polysaccharides. Thus I aim to first use the existing parameter libraries for dextran and cellulose. In case my measurements prove fruitful, I believe I might try to obtain a simple AA-mapped parametrization for amylose. I expect such description might suffice for shorter chains in a diluted system, where the role of helicality is less profound. Ideally, I'd also be compelled to introduce a structural bias to ensure the helical structure, however, I would probably need some additional help in order achieve that.

I plan to continue following your work on polyply and carbohydrates. In case I manage to obtain some parameters on my own, I will keep you informed.

Cheers, Jaka

[JakaBrklanje]

Hi again, I'm returning to this topic somewhat (perhaps it could be moved to discussions, haven't realized that I posted in "Issues").

My goal is now to use an AA level system and then use it as a base to, hopefully, construct a fully functional Martini 3 parameter library for several polysaccharides. Therefore, I'm curious if one could use Polyply to generate short oligo chains even in all-atom description (preferably Glycam or Charmm). If not, what approach did you take to obtain the initial structures and parameterization?

With regards, Jaka

[fgrunewald]

Hi @JakaBrklanje,

Don't worry about the issue vs discussion thing too much. I'll transfer it when I see fit.

To answer your question: In principle yes you can do it. However, it requires some work to implement all the parameters. In order to generate an all-atom reference model for Martini you are probably quicker using CHARMM-GUI or GlycanWeb provided they have the parameters. If they don't we can implement the CHARMM version into polyply. I can provide some help and guidence. I don't know AMBER well enough to feel confident in deriving consistent parameters.

Please let me know if you cannot find Amylose in the mentioned webpackages or if they fail.

Cheers, Fabian

[JakaBrklanje]

Hi Fabian, thank you for the instructions, I will try out CHARMM and GlycanWeb software to produced the topologies then. I actually do have some parametrization in OPLSAA from some previous project, but only plan to use that for comparison (or perhaps, as a fallback in case other my other attempts don't meet the expectations).

Cheers, Jaka

[JakaBrklanje]

Hi again, I am not sure whether the following inquiry is still on topic for this thread, but since I could not find any better communication channel, I'm proceeding here. I was wondering if you use some particular protocol to equilibrate carbohydrate structures generated by Polyply in CG water. So far, all of my attempts to simulate CH oligomers with Martini 3 force field ended up in a state where there is barely any detectable translatory movement of particles, even though the system's temperature is about 300 K. I have tried several approaches, but to no avail. I am aware that Martini 3 should in principle eliminate the need to use the antifreeze solvent particles, however, I'm afraid I can't make it work appropriately. Cheers, Jaka

[JakaBrklanje]

Just as a follow-up - I resolved the issue by carefully adjusting the parameters in insane (py) script. I could not exactly elucidate the cause of freezing, but apparently my system was too densely packed.

[fgrunewald]

@JakaBrklanje thanks for the follow-up. Martini3 water does not freeze at room temperature. Usually, when it does freeze it means the mpd parameters are set incorrectly, there is an excessive amount of salt in the system, or parameters are missing from the bead type definitions. Even when starting from a perfect crystal under NpT conditions it should melt.