number of simulations visited macrostate

[noeliaferruz]

Hi,

Question. I have a model whose macro 3 was only visited by sims 26, 27, and 31.

lag_time = 75
nr_macrostates = 4
model.markovModel(lag_time, nr_macrostates)

2016-09-26 11:07:36,782 - htmd.model - INFO - 99.8% of the data was used
2016-09-26 11:07:36,792 - htmd.model - INFO - Number of trajectories that visited each macrostate:
2016-09-26 11:07:36,793 - htmd.model - INFO - [33  1 43  3]
2016-09-26 11:07:36,794 - htmd.model - INFO - Take care! Macro 1 has been visited only in 1 trajectories:
2016-09-26 11:07:36,795 - htmd.model - INFO - id = 40
parent = None
input = []
trajectory = ['./filtered/e9s3_e7s1p0f3103/output.filtered.xtc']
molfile = ./filtered/filtered.pdb
2016-09-26 11:07:36,796 - htmd.model - INFO - Take care! Macro 3 has been visited only in 3 trajectories:
2016-09-26 11:07:36,797 - htmd.model - INFO - id = 26
parent = None
input = []
trajectory = ['./filtered/e7s1_e1s4p0f1428/output.filtered.xtc']
molfile = ./filtered/filtered.pdb
2016-09-26 11:07:36,798 - htmd.model - INFO - id = 27
parent = None
input = []
trajectory = ['./filtered/e7s2_e2s4p0f2624/output.filtered.xtc']
molfile = ./filtered/filtered.pdb
2016-09-26 11:07:36,799 - htmd.model - INFO - id = 31
parent = None
input = []
trajectory = ['./filtered/e8s1_e1s10p0f2257/output.filtered.xtc']

But then sampleStates shows IDs from sim 38 -and I have also seen id 40-:

 absframes,relframes=model.sampleStates(range(4), [10,10,10,10])
print(relframes[3], absframes[3])

[[  27  461]
 [  38 1950]
 [  38  483]
 [  27  690]
 [  26 1541]
 [  26 3829]
 [  26 1848]
 [  26 3684]
 [  27  433]
 [  31 4891]] [105720 148091 146624 105949 101800 104088 102107 103943 105692 125598]

Hints?

[stefdoerr]

That should definitely not happen. I will try to recreate this on some dataset although it might be tricky. Can you maybe save the model and send it to me?

[noeliaferruz]

mmmh no I don't think so. Thanks.

[j3mdamas]

Is it something I can help with?

[stefdoerr]

If you really want to. Make a Markov model analysis and see if this happens to you.

[j3mdamas]

I mean, if it is some private data that I can look into

[giadefa]

Stefan will soon be able to look at private data as well.

On 27 September 2016 at 10:33, João M. Damas notifications@github.com wrote:

I mean, if it is some private data that I can look into

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/Acellera/htmd/issues/143#issuecomment-249801273, or mute the thread https://github.com/notifications/unsubscribe-auth/AHaqOkwJd2IWx1n3ynJpMOBu7CSI1awdks5quNTvgaJpZM4KGx9P .

http://www.acellera.com

   <https://twitter.com/acellera>

https://www.youtube.com/user/acelleracom https://www.linkedin.com/company/2133167?trk=tyah&trkInfo=clickedVertical%3Acompany%2CclickedEntityId%3A2133167%2Cidx%3A2-1-2%2CtarId%3A1448018583204%2Ctas%3Aacellera https://www.acellera.com/md-simulation-blog-news/ http://is.gd/1eXkbS

[j3mdamas]

@noeliaferruz, is this still an issue?

[j3mdamas]

What I really meant is: can you provide us the info to check this eventual bug?

[noeliaferruz]

Hi @j3mdamas No, Unfortunately, I won´t be able to provide an example. Perhaps a system with very little data would help debug.

[j3mdamas]

@stefdoerr, if it's unreproducible, I would consider closing this issue. What do you think?

[stefdoerr]

Did you try to find an example which has few trajectories in a macrostate? I might look into it

[j3mdamas]

I have no example

[stefdoerr]

@noeliaferruz check the definition of sampleStates specifically argument statetype :smile: But you have a point. The default should be 'macro' and not 'micro'. Changing it now.