Coloc2: calculating the approximate PSF

Hi Stephanmg,

I am also analizing some images using Coloc2 but I don’t know how calculate the approximate PSF. It would be nice if you could tell me how can I do it.

Thank you very much.

Best
Alfonsa

@adiaz

This is info I got from the Coloc2 expert @chalkie666… If images are collected properly the PSF should be close to 3 pixels. To be sure, you can calculate it knowing the optical resolution d = lambda / 2*NA. (NA is the second number written on the objective lens, after the slash after the magnification number.) Then divide d by the pixel spacing in the image and that is the approximate PSF size.

You could also use this tool from SVI: Nyquist Calculator… as ultimately, the psf depends on many factors: the magnification, objective lens, camera binning, immersion oil, mounting medium, …

eta :slight_smile:

Hello Etarena,

Thank you so much for the info. I still have some questions, it would be nice if you could help me:

  1. Is lambda the value of the laser wavelength???
  2. If I use the Nyquist Calculator I do not obtain a value for the PSF, I obtain 3 values in X, Y and Z, Which of these values ​​is the one that I have to use as the value of the PSF in coloc2?

Thank you very much and sorry for the inconveniences,

Best
Alfonsa

Hello Etarena,

This is Alfonsa again, I forgot asked you if the “pixel spacing” is the “pixel size”…

Thank tou so much
Best
Alfonsa

Good day,

there is no pixel size but only a pixel spacing however, because this fact appears to be a big secret, most often both expressions are used in the same sense.

Regards

Herbie

1 Like

Hi Herbie,

Thank you!!!

Best,
Alfonsa

@adiaz

Yes. :slight_smile:

What are the values that you calculate in the end? Can you provide me with an example image so I can take a look myself (make sure it’s in the original formate)?

If you are worried - you can just use @chalkie666’s calculation and you will be fine. :slight_smile: But the size that you put in for the PSF is in pixels… so keep that in mind.

eta

Good morning Etarena,

I’m sorry to bother you and I thank you very much for your help.
I attach the image of a bead, of which I am trying to calculate the value of the PSF to put it in the “PSF” section of the plugin Coloc2.
I did:
d = lambda / 2*NA;
d=488/2(1.4)=174.2;
Then I divided d by the pixel spacing (0.041um = 41nm);
d/pixel spacing = 174.2/41 = 4.2.
PSF = 4.2

Using the Nyquist Calculator I obtain 3 values


as you can observe in this image I obtain 3 values x: 43nm, y: 43nm, z: 130nm; and I don’t know which value is the PSF value I need to Coloc 2.

thank you very much for helping me and sorry for the inconvenience.

Best
Alfonsa

here goes the bead image (It looks like I forgot sending it in my previous post)
Thanks!!

@adiaz

Looks like your magic number is 4 !! :slight_smile:

eta

1 Like

Hello all,

yes I used this calculator too.

Best,
S.

Just to be sure, Coloc2 says:
PSF has to be provided in pixels of my image, correct?

The Nyquist calculator gives in nm:
x: 247 nm
y: 247 nm
z: 3889 nm

Thus if I know my pixel spacing d I can divide by d resulting in the value I’ve to put into Coloc2.

Best wishes,
Stephan

Dear All,

@etadobson

what is actually done with the number you enter in the field PSF in Coloc2?

And why is the default 1.0 not 1?

If the pixel size is larger than the PSF, do I enter:

0.3 or 1 ?? *

If 1 (instead of 0.3) what impact will this have on the results?

(* image again with better sampling is no option in this case)

Thanks &
Kind regards,

Tobias

@Tobias -

I these cases… the best thing to do is to check the source code:

The psf is used in Costes calculations:

1 Like

Dear Eta,

thanks a lot,
as “non java native” it seems to me it should be an integer.

public CostesSignificanceTest(PearsonsCorrelation<T> pc, int psfRadiusInPixels,

Then 1 would be the right answer to my question and the default in Fiji should be 1 not
1.0 (which is a bit confusing)…

Or?

From there on it seems it is just used to calculate the block size.
Thus, 1 instead of 0.3 will not make a difference.

Thanks again & Kind regards

Tobias

1 Like