Question on Dataset Reconstruction

[Jenny Higgins]

Hi all,

I am new to the wonderful world of super-resolution microscopy. I have
been looking at the QuickPALM code to try to better understand how it
works, in particular with how it handles error during reconstruction
from a dataset. I noticed that the dataset contains a left, right, up,
and down width/height, but I am having trouble figuring out if those
are taken into account during image reconstruction or not.

In summary, is error taken into account during image reconstruction,
and if so, how?

Any help would be greatly appreciated. Thank you for your time!

-Jenny Higgins

[Ricardo Henriques]

Hi Jenny,

I am new to the wonderful world of super-resolution microscopy.

Welcome :).

 

I have been looking at the QuickPALM code to try to better understand how it
works, in particular with how it handles error during reconstruction
from a dataset. I noticed that the dataset contains a left, right, up,
and down width/height, but I am having trouble figuring out if those
are taken into account during image reconstruction or not.

Do you mean localization accuracy (error) or software exceptions (error)?

I'll assume the first. 

Width and height parameters are used for two main things - to calculate the position of the particles in Z if you're using astigmatism and as a quality control step, on this latter case they are used to calculate the symmetry of the particle and particles that do not have a minimum symmetry as given by the threshold you impose on the particle detection dialog are discarded.

The reconstruction is composed by the particles that are on the particle table and that have passed the following three tests: intensity needs to be high enough (the SNR threshold), the size needs to be similar to the given maximum full-width-half-maximum and it needs to have a high enough symmetry.

In summary, is error taken into account during image reconstruction,
and if so, how?

In this case, the error is dealt by simple test to verify if the particle has expected properties. But you can evolve this by directly looking into the properties of the collected particles, for example, in theory the localization accuracy is proportional to the inverse of the square-root of the particle intensity [LA = X / sqrt(N), where X is related to the size of your systems PSF and N the number of photons]. 

 

Any help would be greatly appreciated. Thank you for your time!

Well I hope this helped somehow, also check out the sup. mat. of the QuickPALM paper - I can send it to you if you don't have access.

Let me know if you have further questions.

 

Best regards,

Ricardo

--
Ricardo Henriques
PhD Student
Gene Expression and Biophysics Unit
Institute of Molecular Medicine
Faculty of Medicine, University of Lisbon
Av. Prof. Egas Moniz
1649-028 Lisbon, Portugal
Phone: + 351 217999503, Ext: 47318
Fax: + 351 217999504
E-mail: rhenr...@fm.ul.pt