down/ and up/ directories

Hey!

I divided my membrane into 60 bins and after I run umbrella sampling simulations using Gromacs, I have 60 different umbrella{i}_pullf.xvg and umbrella{i}_pullx.xvg (i= 0 to 50) files. Do I need to put these files in two down/ and up/ directories and then run the setup.sh and getDall.py to get the diffusion curve?
Actually I did and the result(attached photo) was nonsense.

I would be appreciated if you can advise me to fix it.

[cnrowley]

How frequently are you saving your steps from your simulation? For diffusion, you usually have to save them at every step.

…

On Tue, Sep 15, 2020 at 7:45 AM Javad ***@***.***> wrote: Hey! I divided my membrane into 60 bins and after I run umbrella sampling simulations, I have 6 [image: Screenshot 2020-09-15 at 6 15 12 PM] <https://user-images.githubusercontent.com/42797415/93198308-6d9af300-f77f-11ea-92c4-952ffeecf3c3.png> 0 different umbrella{i}_pullf.xvg and umbrella{i}_pullx.xvg (i= 0 to 50) files. Do I need to put these files in two down/ and up/ directories and then run the setup.sh and getDall.py to get the diffusion curve? Actually I did and the result(attached photo) was nonsense. I would be appreciated if you can advise me to fix it. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABYZVJBYIUJRNF3U2Z74ZOTSF45F7ANCNFSM4RM4R2QA> .

Thanks for your reply.

I just set -s value to 2 as it was in the original script. I have run every bin of membrane for 50 ns so it means I should set the value of s to 25000000?
Also, do you know what's the optimum value for --maxcorr ? it looks changing its value, doesn't change the curve for my simulation

[cnrowley]

No, how often are you saving the position of the solute in your MD simulation?

…

On Tue, Sep 15, 2020 at 12:02 PM Javad ***@***.***> wrote: Thanks for your reply. Do you mean --maxcorr value? I set it 10000 in setup.sh script. Also, I have run every bin of membrane for 50 ns. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABYZVJFM3FWUHSYJK62OYXTSF53IRANCNFSM4RM4R2QA> .

I have set nstxout-compressed = 500 in my simulation which is 50ns run time for every bin (60 bins in total).
Can you advise me on how I should set the input values to get the right diffusion curve?

Based on my understanding, I should set -f value to 1 as I used Gromacs, -s value to 2000 as every step in my simulation is 2 ps which is 2000 fs and --maxcorr to 50000 , but I face below error for every bin:

#Time step of 2000 fs will be used.
#No cutoff will be used to integrate the ACF.
#D(s) will be saved in out_up_sim_2.out
#var 1.15969e+10
#varVel 6.54981e-30
#WARNING: ACF has only decayed to 94.4956% of initial value at the end of the region of integration.
libc++abi.dylib: terminating with uncaught exception of type boost::wrapexceptstd::domain_error: Error in function boost::math::tools::toms748_solve: Parameters a and b do not bracket the root: a=9.9999999999999995e-07
setup.sh: line 6: 34857 Abort trap: 6 /Users/javad/Downloads/ACFCalculator-master/build/ACFcalculator -i umbrella${i}pullx.xvg -t gromacs -a acf_up_sim${i}.dat -o out_up_sim_${i}.out -f 1 -s 2000 --maxcorr 50000

[cnrowley]

This is a somewhat different issue. You have to save a time series for the coordinate you are calculating the diffusion coefficient for at every step. I would recommend reading the paper https://pubs.acs.org/doi/abs/10.1021/acs.jctc.6b00747 carefully before attempting to do these calculations.

…

On Wed, Sep 16, 2020 at 12:48 AM Javad ***@***.***> wrote: I have set nstxout-compressed = 500 in my simulation. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#1 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABYZVJB3IJJL2BGJGVKZROLSGAVCDANCNFSM4RM4R2QA> .