Gmx order - is it possible to get standard deviation/error of the results?

User discussions
1

GROMACS version: 2022
GROMACS modification: No
I have a question. Is it possible to get errors, for example standard deviation of my results in gmx order?
For MGDG lipid I have results like:

2 0.193571
3 0.303616
4 0.295035
5 0.301556
6 0.277012
7 0.270018
8 0.216542
9 0.207125
10 0.0234358
11 -0.0647441
12 0.166839
13 0.0761101
14 0.0373471
15 0.112092
16 0.0335867
17 0.0100009

I thought about using -permolecule option and then get SCD of each molecule, calculate deviation and average. Probably I could also use gmx order every 10 ps in for loop in bash script and then calculate deviation and average, but It would take a lot of time (for 100 ns I need to run 10000 times gmx order) and processing of files would be laborious, but I would have a standard deviation with respect to molecules and frames.

I wonder if there is an easier way?

2

Subdividing the trajectory over time is probably the best bet, but an interval of 10 ps is way too short. You probably shouldn’t even be saving frames that often for a system like this. Relaxation times of lipids are much longer than the ps scale. If you have 100 ns, maybe break that into 10-ns chunks of time, but I wouldn’t go much shorter than that.

1 Like