Running Backwards on Large Lipid-Only Systems
- naveen
-
Topic Author
- Visitor
6 years 1 week ago #7620
by naveen
Running Backwards on Large Lipid-Only Systems was created by naveen
Hi all,
I'm currently trying to use Backward to map from Martini to Gromos53a6 on a large lipid system. My lipid system is a 4 Component System comprised of DBPC, PIPC, DIPC, CHOL with about ~4,500 total lipids.
I have gotten Backwards to work correctly on single lipid gro files for all my lipids, and even on some small 64/64 bilayers. However, when I run a large system, my system always "blows up" within about 0.1 ps.
Has anyone run Backwards on large systems similar to mine? What are some possible reasons small bilayers/single lipids would work correctly, but a large system woud not?
Thanks,
Naveen
I'm currently trying to use Backward to map from Martini to Gromos53a6 on a large lipid system. My lipid system is a 4 Component System comprised of DBPC, PIPC, DIPC, CHOL with about ~4,500 total lipids.
I have gotten Backwards to work correctly on single lipid gro files for all my lipids, and even on some small 64/64 bilayers. However, when I run a large system, my system always "blows up" within about 0.1 ps.
Has anyone run Backwards on large systems similar to mine? What are some possible reasons small bilayers/single lipids would work correctly, but a large system woud not?
Thanks,
Naveen
Please Log in or Create an account to join the conversation.
- riccardo
-
- Offline
- Platinum Boarder
Less
More
6 years 1 week ago - 6 years 1 week ago #7621
by riccardo
Replied by riccardo on topic Running Backwards on Large Lipid-Only Systems
Hi Naveen,
Does the small lipid patch you backmapped successfully contain all the lipids + chol that you have in the bigger system?
If so, the problem may just be that the size of the system requires longer minimization steps. Do the minimizations converge or there are still some high forces at the end of the 2nd minimization? In the latter case, did you already try with the flags -em and -nb to make the 2 minimizations longer (both are 500 steps long by default, but you could increase them to -em 1500 and -nb 1500 for example and see if that changes things)?
Does the small lipid patch you backmapped successfully contain all the lipids + chol that you have in the bigger system?
If so, the problem may just be that the size of the system requires longer minimization steps. Do the minimizations converge or there are still some high forces at the end of the 2nd minimization? In the latter case, did you already try with the flags -em and -nb to make the 2 minimizations longer (both are 500 steps long by default, but you could increase them to -em 1500 and -nb 1500 for example and see if that changes things)?
Last edit: 6 years 1 week ago by riccardo.
Please Log in or Create an account to join the conversation.
- naveen
-
Topic Author
- Visitor
6 years 6 days ago #7624
by naveen
Replied by naveen on topic Running Backwards on Large Lipid-Only Systems
I have increased the number of EM and NB steps. The EM always converges at Force~1E4. Is this why my simulation is failing later? Because my Energy is too high?What Force should I target for my EM?
I will try a smaller patch as well.
I will try a smaller patch as well.
Please Log in or Create an account to join the conversation.
- riccardo
-
- Offline
- Platinum Boarder
6 years 4 days ago #7625
by riccardo
Replied by riccardo on topic Running Backwards on Large Lipid-Only Systems
A force of ~1E4 at the end of the EM could be okay. How about the second (NB) minimization? That should give you a final negative potential energy (~ 1E^(-7) or lower I'd guess). If the NB does not go to negative potential energies because of too high forces, you could check on which atom the high force is acting: the output of the minimization prints out the atom on which the highest force recorded is acting. You could then visually inspect your system to check what's causing that high force on that atom.
Please Log in or Create an account to join the conversation.
Time to create page: 0.092 seconds