Closed mr-september closed 2 years ago
Not sure if I am on the right track, but I did notice that the mmsd
object generated from mmsd = ma.getMeanMSD
contains only Inf
or NaN
values.
In the getMeanMSD.m
, lines 109-110 appear to be relevant:
% Standard deviation
mstd = sqrt( sum_weight ./ (sum_weight.^2 - sum_square_weight) .* sum_weighted_variance );
After some more thought on this, I think it's a misunderstanding on my part on what is being done.
Typically, there are multiple particles/spots and tracks, from which statistics like mean, std deviation arise.
In my case, since only 1 particle is tracked, inf/NaN values should be expected for std deviation.
In a similar vein, it also makes no sense to "fit" a graph onto a single sampled curve.
Assuming this misunderstanding is the cause, I will close this issue. Please re-open if I am wrong!
Hi, I have been trying to do MSD analysis, following parts of this manual and this tutorial. However, I am hitting a wall with NaN values:
My whole Matlab script (the error comes from the very last section
% calculate diffusion coefficiency
):And my
PEI_chan00_Tracks.xml
as outputted from TrackMate on FIJI (sorry about the mess, GitHub doesn't allow attaching this format):Apologies in advance because this sounds like a rather trivial problem on the surface, but I am quite new to Matlab, and I have no clue how to even begin looking for these NaN values.