# What is the good way to extract the representative PDB structure from 300 ns simulation

**URL:** <https://gromacs.bioexcel.eu/t/what-is-the-good-way-to-extract-the-representative-pdb-structure-from-300-ns-simulation/1934>\
**Category:** User discussions\
**Created:** [April 10, 2021, 12:02am UTC](https://gromacs.bioexcel.eu/t/what-is-the-good-way-to-extract-the-representative-pdb-structure-from-300-ns-simulation/1934 "2021-04-10T00:02:20Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![minghao2121](https://avatars.discourse-cdn.com/v4/letter/m/c67d28/32.png) [@minghao2121](https://gromacs.bioexcel.eu/u/minghao2121)\
**Post date:** [April 10, 2021, 12:02am UTC](https://gromacs.bioexcel.eu/t/what-is-the-good-way-to-extract-the-representative-pdb-structure-from-300-ns-simulation/1934/1 "2021-04-10T00:02:20Z")

</div>

GROMACS version: 2021.1  
GROMACS modification: Yes  
Hello,  
I have obtained multiple 300-ns simulation and did some following analysis such as RMSF.  
Now I want to extract the representative PDB structure from this 300-ns simulation.

I know we can use ‘gmx cluster’ ([gmx cluster — GROMACS 2021 documentation](https://manual.gromacs.org/documentation/2021/onlinehelp/gmx-cluster.html)) to do this.

Now the question is that: there are too many structures in the \*.xtc file because I use the md.mdp configured with (dt = 0.002 ; 2 fs nstxout-compressed = 5000 ; save compressed coordinates every 10.0 ps). How can I use -tu and -b and -e options for just clustering those structures saying from 200ns - 300ns?

Or I must do clustering for the while 300-ns \*.xtc file?

For the representative structure, be should it from the last saying 100-ns **stable** structure? must from the stable structures (checking by looking at the RMSD vs. time (ns), where supposing the last 100-ns give the stable line meaning no large fluctuates.)

Or can you give some suggestions how can I get the proper representative PDB structure for further analysis (e.g., comparing this wild-type structure with the structure from the mutation-based same 300-ns simulation)?

Thanks,  
Ming

---

<div class="post-metadata">

**Author:** ![alevilla](https://dub1.discourse-cdn.com/flex017/user_avatar/gromacs.bioexcel.eu/alevilla/32/439_2.png) [@alevilla](https://gromacs.bioexcel.eu/u/alevilla)\
**Post date:** [April 12, 2021, 9:18am UTC](https://gromacs.bioexcel.eu/t/what-is-the-good-way-to-extract-the-representative-pdb-structure-from-300-ns-simulation/1934/2 "2021-04-12T09:18:10Z")

</div>

Hi,

> [@minghao2121](#):
>
> Now the question is that: there are too many structures in the \*.xtc file because I use the md.mdp configured with (dt = 0.002 ; 2 fs nstxout-compressed = 5000 ; save compressed coordinates every 10.0 ps). How can I use -tu and -b and -e options for just clustering those structures saying from 200ns - 300ns?

you can try gmx cluster -with the option -tu ns -b 200 -e 300

or in alternative you can use -dt option to reduce the number of frame to analyze

best regards  
Alessandra

---

<div class="post-metadata">

**Author:** ![minghao2121](https://avatars.discourse-cdn.com/v4/letter/m/c67d28/32.png) [@minghao2121](https://gromacs.bioexcel.eu/u/minghao2121)\
**Post date:** [April 12, 2021, 3:13pm UTC](https://gromacs.bioexcel.eu/t/what-is-the-good-way-to-extract-the-representative-pdb-structure-from-300-ns-simulation/1934/3 "2021-04-12T15:13:09Z")

</div>

> [@alevilla](#):
>
> Alessandra

Hi Alessandra,

Thanks for your help. I will try that.

Best,  
Ming