Hello,
when we used vbSPT software in our lab, we faced a problem of finding extra states. For example, when we do not expect more than 2 physical states, vbSPT can find 3 or 4 states sometimes. I say 'sometimes' because I get different number of states when I analyse images from different fields of view of one sample. I think it is an indication that not all of the states are actually physical. Also the so-called 'extra' states I get in my analysis have a really small occupancy (compared to the other states) and close values of diffusion coefficients to one of the 'physical' states.
My questions are: Have you noticed it analyzing your data?
Do you know a good way how to automatically distinguish 'physical' from 'extra' states?
Cheers,
Oleg
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Note to readers: this discussion moved offline to allow exchange of private data. Below is a brief summary.
Dear Oleg,
From the example data you sent privately, my guess is that you are affected by confinement effects, since your bacteria are long and thin, and you are following a small protein that can be expected to have a high (~ 10 um^2/s) diffusion constant.
The confinement is only in the transverse (radial) direction though, and so to minimize confinement artifacts, I would recommend you preprocess your data similar to how we did for E coli in the Nat. Methods paper, namely
1. rotate the coordinates of each bacteria so that the x-axis is aligned with the bacterial long axis, and
2. run an 1-dimensional analysis on the motion along the (rotated) x-coordinates only.
Best,
Martin
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear Martin,
Thanks for answering and solving my last question regarding 'arrow.m'
This time I was trying to use 'visualize_localisation.m' and it gave me the following error message:
visualize_localisation
Read options structure based on runinput file runinput_normal.m
Undefined variable "hmm" or class "hmm.Wbest".
Error in visualize_localisation (line 14)
W=hmm.Wbest;
Please Help me!!
Thanks and regards
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
when we used vbSPT software in our lab, we faced a problem of finding extra states. For example, when we do not expect more than 2 physical states, vbSPT can find 3 or 4 states sometimes. I say 'sometimes' because I get different number of states when I analyse images from different fields of view of one sample. I think it is an indication that not all of the states are actually physical. Also the so-called 'extra' states I get in my analysis have a really small occupancy (compared to the other states) and close values of diffusion coefficients to one of the 'physical' states.
My questions are: Have you noticed it analyzing your data?
Do you know a good way how to automatically distinguish 'physical' from 'extra' states?
Cheers,
Oleg
Note to readers: this discussion moved offline to allow exchange of private data. Below is a brief summary.
Dear Oleg,
From the example data you sent privately, my guess is that you are affected by confinement effects, since your bacteria are long and thin, and you are following a small protein that can be expected to have a high (~ 10 um^2/s) diffusion constant.
The confinement is only in the transverse (radial) direction though, and so to minimize confinement artifacts, I would recommend you preprocess your data similar to how we did for E coli in the Nat. Methods paper, namely
1. rotate the coordinates of each bacteria so that the x-axis is aligned with the bacterial long axis, and
2. run an 1-dimensional analysis on the motion along the (rotated) x-coordinates only.
Best,
Martin
When I tried to run the program:
Undefined operator '*' for input arguments of type 'matlab.graphics.axis.Axes'.
Error in arrow (line 423)
ax = 0 * gca;
Error in VB3_displayHMMmodel (line 199)
arrow(stateCoord1, stateCoord2, 'wid', round(transitions(1, 2)*100));
Please Help, I am using MATLAB 2016b and 2019b
Dear Alok,
I think the problem might be that arrow.m is an old version that is not compatible with your matlab versions, see https://se.mathworks.com/matlabcentral/answers/230792-why-does-arrow-fail-r2015a
Please try to update Tools/displayHMMmodel/private/arrow.m in vbSPT with the latest version from https://se.mathworks.com/matlabcentral/fileexchange/278-arrow and see if that helps.
Best regards,
Martin
Dear Martin,
Thanks for answering and solving my last question regarding 'arrow.m'
This time I was trying to use 'visualize_localisation.m' and it gave me the following error message:
Error in visualize_localisation (line 14)
W=hmm.Wbest;
Please Help me!!
Thanks and regards