discussion is open...continues ... about localization and super resolution

[Alberto Diaspro]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

http://www.nature.com/nmeth/journal/v11/n3/full/nmeth.2843.html

[lechristophe]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

This is a great introduction to a question a wanted to ask to the SMLM
specialists (I did not find the answer in the above reference). I also
posted it on the Nanoscopy LinkedIn group, so I'm sorry for people that
will get it two times!

There are several formulas to estimate the lateral (XY) precision of
individual localization events in PALM/STORM microscopy. This goes from the
crude sigma(PSF)/√Nphotons to more elaborate formulas that includes
pixelation and background (such as Thomson & Webb Biophys J 2002, corrected
in Mortensen Nat Methods 2010). What I'm looking for is an equation that
allows to estimate the precision for the Z localization when using
astigmatism-based 3D PALM/STORM. The reference I have found where this
precision is studied theoretically (such as Badierostami Applied Phys Lett
2010, Shaevitz Int J Optics 2009) do not use a directly calculable formula.

I assume the precision depends on the photon number and the point on the Z
calibration curve with the respective X and Y PSF width. Is anyone aware of
such a formula?

Best Regards,

Christophe




2014-03-04 21:52 GMT+01:00 Alberto Diaspro <[hidden email]>:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> http://www.nature.com/nmeth/journal/v11/n3/full/nmeth.2843.html
>

[Sripad Ram]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Chris,
The z-localization precision for astigmatism based 3D localization was worked out by holtzer et al., applied physics letters, 2007, 90, 053902.

On a related note, the precise calculations for the X/Y localization precision that includes background and pixilation was published in 2004 by Raimund Ober group (Ober et al; Biophysical Journal, 2004, 86:1185-1200). You can download the software to calculate the localization precision from here  (http://www.wardoberlab.com/software/fandplimittool/).

Hope this helps.

Regards,
Sripad


-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Christophe Leterrier
Sent: Tuesday, March 04, 2014 3:16 PM
To: [hidden email]
Subject: Re: discussion is open...continues ... about localization and super resolution

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

This is a great introduction to a question a wanted to ask to the SMLM specialists (I did not find the answer in the above reference). I also posted it on the Nanoscopy LinkedIn group, so I'm sorry for people that will get it two times!

There are several formulas to estimate the lateral (XY) precision of individual localization events in PALM/STORM microscopy. This goes from the crude sigma(PSF)/√Nphotons to more elaborate formulas that includes pixelation and background (such as Thomson & Webb Biophys J 2002, corrected in Mortensen Nat Methods 2010). What I'm looking for is an equation that allows to estimate the precision for the Z localization when using astigmatism-based 3D PALM/STORM. The reference I have found where this precision is studied theoretically (such as Badierostami Applied Phys Lett 2010, Shaevitz Int J Optics 2009) do not use a directly calculable formula.

I assume the precision depends on the photon number and the point on the Z calibration curve with the respective X and Y PSF width. Is anyone aware of such a formula?

Best Regards,

Christophe




2014-03-04 21:52 GMT+01:00 Alberto Diaspro <[hidden email]>:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> http://www.nature.com/nmeth/journal/v11/n3/full/nmeth.2843.html
>

[Zdenek Svindrych]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi,

also a brute force simulation of many images of many molecules and
subsequent localisation with your preferred method is a straightforward
(albeit a bit time consuming) way to get the localisation accuracy (and
precision...) at nearly real-life conditions. You can simulate any photon
counts, background, detector pixelation, effect of molecule density, biases
of the localisation algos and so on.

Hopefully our group (Guy Hagen et al) will soon publish a software tool to
greatly simplify this. But I still didn't find a satisfactory astigmatic PSF
model to include into the simulation. The cylindrical gaussian is really not
close approximation of our experimental results...

Best,

zdenek svindrych
lf1.cuni.cz



---------- Původní zpráva ----------
Od: Sripad Ram <[hidden email]>
Komu: [hidden email]
Datum: 4. 3. 2014 23:17:16
Předmět: Re: discussion is open...continues ... about localization and super
resolution

"*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Chris,
The z-localization precision for astigmatism based 3D localization was
worked out by holtzer et al., applied physics letters, 2007, 90, 053902.

On a related note, the precise calculations for the X/Y localization
precision that includes background and pixilation was published in 2004 by
Raimund Ober group (Ober et al; Biophysical Journal, 2004, 86:1185-1200).
You can download the software to calculate the localization precision from
here (http://www.wardoberlab.com/software/fandplimittool/).

Hope this helps.

Regards,
Sripad


-----Original Message-----
From: Confocal Microscopy List [mailto:[hidden email]] On
Behalf Of Christophe Leterrier
Sent: Tuesday, March 04, 2014 3:16 PM
To: [hidden email]
Subject: Re: discussion is open...continues ... about localization and super
resolution

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

This is a great introduction to a question a wanted to ask to the SMLM
specialists (I did not find the answer in the above reference). I also
posted it on the Nanoscopy LinkedIn group, so I'm sorry for people that will
get it two times!

There are several formulas to estimate the lateral (XY) precision of
individual localization events in PALM/STORM microscopy. This goes from the
crude sigma(PSF)/√Nphotons to more elaborate formulas that includes
pixelation and background (such as Thomson & Webb Biophys J 2002, corrected
in Mortensen Nat Methods 2010). What I'm looking for is an equation that
allows to estimate the precision for the Z localization when using
astigmatism-based 3D PALM/STORM. The reference I have found where this
precision is studied theoretically (such as Badierostami Applied Phys Lett
2010, Shaevitz Int J Optics 2009) do not use a directly calculable formula.

I assume the precision depends on the photon number and the point on the Z
calibration curve with the respective X and Y PSF width. Is anyone aware of
such a formula?

Best Regards,

Christophe




2014-03-04 21:52 GMT+01:00 Alberto Diaspro <[hidden email]>:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> http://www.nature.com/nmeth/journal/v11/n3/full/nmeth.2843.html
>"

[lechristophe]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Thank you for all these informative answers!

The Holtzer et al. article is providing a formula for Z-loclaization
precision in the astigmatism case (although they have a x2 factor between
theory and simulation because they don't explicitely take pixelation into
account).

I also found this recent article:

Rieger B, Stallinga S. The Lateral and Axial Localization Uncertainty in
Super-Resolution Light Microscopy. Chemphyschem. 2013 Dec 2. doi:
10.1002/cphc.201300711. [Epub ahead of print] PubMed PMID: 24302478.

http://onlinelibrary.wiley.com/doi/10.1002/cphc.201300711/abstract

Here they explicitly approximate the Z-localization precision of biplane,
astigmatism and double-helix modalities, taking into account pixelation and
background with a claimed <1% error. Although I don't really understand the
formulas yet, because they depend on the "the distance between focal lines"
and "a measure of the focal depth" that are "obtained by a calibration
measurement" but not explicited further it seems.

Thanks again,

Christophe




2014-03-04 23:33 GMT+01:00 Zdenek Svindrych <[hidden email]>:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi,
>
> also a brute force simulation of many images of many molecules and
> subsequent localisation with your preferred method is a straightforward
> (albeit a bit time consuming) way to get the localisation accuracy (and
> precision...) at nearly real-life conditions. You can simulate any photon
> counts, background, detector pixelation, effect of molecule density, biases
> of the localisation algos and so on.
>
> Hopefully our group (Guy Hagen et al) will soon publish a software tool to
> greatly simplify this. But I still didn't find a satisfactory astigmatic
> PSF
> model to include into the simulation. The cylindrical gaussian is really
> not
> close approximation of our experimental results...
>
> Best,
>
> zdenek svindrych
> lf1.cuni.cz
>
>
>
> ---------- Původní zpráva ----------
> Od: Sripad Ram <[hidden email]>
> Komu: [hidden email]
> Datum: 4. 3. 2014 23:17:16
> Předmět: Re: discussion is open...continues ... about localization and
> super
> resolution
>
> "*****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi Chris,
> The z-localization precision for astigmatism based 3D localization was
> worked out by holtzer et al., applied physics letters, 2007, 90, 053902.
>
> On a related note, the precise calculations for the X/Y localization
> precision that includes background and pixilation was published in 2004 by
> Raimund Ober group (Ober et al; Biophysical Journal, 2004, 86:1185-1200).
> You can download the software to calculate the localization precision from
> here (http://www.wardoberlab.com/software/fandplimittool/).
>
> Hope this helps.
>
> Regards,
> Sripad
>
>
> -----Original Message-----
> From: Confocal Microscopy List [mailto:[hidden email]]
> On
> Behalf Of Christophe Leterrier
> Sent: Tuesday, March 04, 2014 3:16 PM
> To: [hidden email]
> Subject: Re: discussion is open...continues ... about localization and
> super
> resolution
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> This is a great introduction to a question a wanted to ask to the SMLM
> specialists (I did not find the answer in the above reference). I also
> posted it on the Nanoscopy LinkedIn group, so I'm sorry for people that
> will
> get it two times!
>
> There are several formulas to estimate the lateral (XY) precision of
> individual localization events in PALM/STORM microscopy. This goes from the
> crude sigma(PSF)/√Nphotons to more elaborate formulas that includes
> pixelation and background (such as Thomson & Webb Biophys J 2002, corrected
> in Mortensen Nat Methods 2010). What I'm looking for is an equation that
> allows to estimate the precision for the Z localization when using
> astigmatism-based 3D PALM/STORM. The reference I have found where this
> precision is studied theoretically (such as Badierostami Applied Phys Lett
> 2010, Shaevitz Int J Optics 2009) do not use a directly calculable formula.
>
> I assume the precision depends on the photon number and the point on the Z
> calibration curve with the respective X and Y PSF width. Is anyone aware of
> such a formula?
>
> Best Regards,
>
> Christophe
>
>
>
>
> 2014-03-04 21:52 GMT+01:00 Alberto Diaspro <[hidden email]>:
>
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > *****
> >
> > http://www.nature.com/nmeth/journal/v11/n3/full/nmeth.2843.html
> >"
>

[Zdenek Svindrych]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Christophe,



the strange 'd' and 'l' parameters are the same as in the holtzer et al.
paper mentioned earlier in this thread.




In the 'bi-plane" method the 'l' parameter is the axial distance of the two
planes of focus in the sample space. In the astigmatic method it is the
amount of astigmatism, again referred to the sample space. That is the axial
distance between the best focus in x direction and best focus in  the y
direction.




The 'd' parameter is (most likely) a measure of the depth of field and it
depends on the numerical aperture. But I haven't found any analytic
approximation. They simply got these parameters by fitting experimental
PSFs.




Regards,

zdenek svindrych






---------- Původní zpráva ----------
Od: Christophe Leterrier <[hidden email]>
Komu: [hidden email]
Datum: 5. 3. 2014 10:55:10
Předmět: Re: discussion is open...continues ... ab out localization and
super resolution

"*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Thank you for all these informative answers!

The Holtzer et al. article is providing a formula for Z-loclaization
precision in the astigmatism case (although they have a x2 factor between
theory and simulation because they don't explicitely take pixelation into
account).

I also found this recent article:

Rieger B, Stallinga S. The Lateral and Axial Localization Uncertainty in
Super-Resolution Light Microscopy. Chemphyschem. 2013 Dec 2. doi:
10.1002/cphc.201300711. [Epub ahead of print] PubMed PMID: 24302478.

http://onlinelibrary.wiley.com/doi/10.1002/cphc.201300711/abstract

Here they explicitly approximate the Z-localization precision of biplane,
astigmatism and double-helix modalities, taking into account pixelation and
background with a claimed <1% error. Although I don't really understand the
formulas yet, because they depend on the "the distance between focal lines"
and "a measure of the focal depth" that are "obtained by a calibration
measurement" but not explicited further it seems.

Thanks again,

Christophe




2014-03-04 23:33 GMT+01:00 Zdenek Svindrych <[hidden email]>:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi,
>
> also a brute force simulation of many images of many molecules and
> subsequent localisation with your preferred method is a straightforward
> (albeit a bit time consuming) way to get the localisation accuracy (and
> precision...) at nearly real-life conditions. You can simulate any photon
> counts, background, detector pixelation, effect of molecule density,

biases

> of the localisation algos and so on.
>
> Hopefully our group (Guy Hagen et al) will soon publish a software tool to
> greatly simplify this. But I still didn't find a satisfactory astigmatic
> PSF
> model to include into the simulation. The cylindrical gaussian is really
> not
> close approximation of our experimental results...
>
> Best,
>
> zdenek svindrych
> lf1.cuni.cz
>
>
>
> ---------- Původní zpráva ----------
> Od: Sripad Ram <[hidden email]>
> Komu: [hidden email]
> Datum: 4. 3. 2014 23:17:16
> Předmět: Re: discussion is open...continues ... about localization and
> super
> resolution
>
> "*****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi Chris,
> The z-localization precision for astigmatism based 3D localization was
> worked out by holtzer et al., applied physics letters, 2007, 90, 053902.
>
> On a related note, the precise calculations for the X/Y localization
> precision that includes background and pixilation was published in 2004 by
> Raimund Ober group (Ober et al; Biophysical Journal, 2004, 86:1185-1200).
> You can download the software to calculate the localization precision from
> here (http://www.wardoberlab.com/software/fandplimittool/).
>
> Hope this helps.
>
> Regards,
> Sripad
>
>
> -----Original Message-----
> From: Confocal Microscopy List [mailto:[hidden email]]
> On
> Behalf Of Christophe Leterrier
> Sent: Tuesday, March 04, 2014 3:16 PM
> To: [hidden email]
> Subject: Re: discussion is open...continues ... about localization and
> super
> resolution
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> This is a great introduction to a question a wanted to ask to the SMLM
> specialists (I did not find the answer in the above reference). I also
> posted it on the Nanoscopy LinkedIn group, so I'm sorry for people that
> will
> get it two times!
>
> There are several formulas to estimate the lateral (XY) precision of
> individual localization events in PALM/STORM microscopy. This goes from

the
> crude sigma(PSF)/√Nphotons to more elaborate formulas that includes
> pixelation and background (such as Thomson & Webb Biophys J 2002,
corrected
> in Mortensen Nat Methods 2010). What I'm looking for is an equation that
> allows to estimate the precision for the Z localization when using
> astigmatism-based 3D PALM/STORM. The reference I have found where this
> precision is studied theoretically (such as Badierostami Applied Phys Lett
> 2010, Shaevitz Int J Optics 2009) do not use a directly calculable
formula.
>
> I assume the precision depends on the photon number and the point on the Z
> calibration curve with the respective X and Y PSF width. Is anyone aware
of

> such a formula?
>
> Best Regards,
>
> Christophe
>
>
>
>
> 2014-03-04 21:52 GMT+01:00 Alberto Diaspro <[hidden email]>:
>
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > *****
> >
> > http://www.nature.com/nmeth/journal/v11/n3/full/nmeth.2843.html
> >"
>"

[Raghu Parthasarathy]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****


Regarding localization accuracy in super-resolution imaging:  Alex Small and I wrote a review recently summarizing and explaining the theory underlying it, and describing its practical implementation in various algorithms: http://www.annualreviews.org/doi/abs/10.1146/annurev-physchem-040513-103735 
Feel free to email me if you can't access the PDF.  ([hidden email])

Relatedly, some 2D tools, with MATLAB code, including simulated images and a GUI for various algorithms (including mine!) are here: http://physics-server.uoregon.edu/~raghu/particle_tracking.html

best wishes,

Raghu


 
--
Raghuveer Parthasarathy
[hidden email]
Group web page: http:&#x2F;&#x2F;physics.uoregon.edu&#x2F;~raghu&#x2F;
Blog: http:&#x2F;&#x2F;eighteenthelephant.wordpress.com&#x2F;


Associate Professor
Department of Physics
1274 University of Oregon
Eugene, OR 97403-1274
http:&#x2F;&#x2F;physics.uoregon.edu&#x2F;~raghu&#x2F;



On Wednesday, March 5, 2014 12:58 AM, Christophe Leterrier <[hidden email]> wrote:
 
*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Thank you for all these informative answers!

The Holtzer et al. article is providing a formula for Z-loclaization
precision in the astigmatism case (although they have a x2 factor between
theory and simulation because they don't explicitely take pixelation into
account).

I also found this recent article:

Rieger B, Stallinga S. The Lateral and Axial Localization Uncertainty in
Super-Resolution Light Microscopy. Chemphyschem. 2013 Dec 2. doi:
10.1002/cphc.201300711. [Epub ahead of print] PubMed PMID: 24302478.

http://onlinelibrary.wiley.com/doi/10.1002/cphc.201300711/abstract

Here they explicitly approximate the Z-localization precision of biplane,
astigmatism and double-helix modalities, taking into account pixelation and
background with a claimed <1% error. Although I don't really understand the
formulas yet, because they depend on the "the distance between focal lines"
and "a measure of the focal depth" that are "obtained by a calibration
measurement" but not explicited further it seems.

Thanks again,

Christophe





2014-03-04 23:33 GMT+01:00 Zdenek Svindrych <[hidden email]>:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi,
>
> also a brute force simulation of many images of many molecules and
> subsequent localisation with your preferred method is a straightforward
> (albeit a bit time consuming) way to get the localisation accuracy (and
> precision...) at nearly real-life conditions. You can simulate any photon
> counts, background, detector pixelation, effect of molecule density, biases
> of the localisation algos and so on.
>
> Hopefully our group (Guy Hagen et al) will soon publish a software tool to
> greatly simplify this. But I still didn't find a satisfactory astigmatic
> PSF
> model to include into the simulation. The cylindrical gaussian is really
> not
> close approximation of our experimental results...
>
> Best,
>
> zdenek svindrych
> lf1.cuni.cz
>
>
>
> ---------- Původní zpráva ----------
> Od: Sripad Ram <[hidden email]>
> Komu: [hidden email]
> Datum: 4. 3. 2014 23:17:16
> Předmět: Re: discussion is open...continues ... about localization and
> super
> resolution
>
> "*****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi Chris,
> The z-localization precision for astigmatism based 3D localization was
> worked out by holtzer et al., applied physics letters, 2007, 90, 053902.
>
> On a related note, the precise calculations for the X/Y localization
> precision that includes background and pixilation was published in 2004 by
> Raimund Ober group (Ober et al; Biophysical Journal, 2004, 86:1185-1200).
> You can download the software to calculate the localization precision from
> here (http://www.wardoberlab.com/software/fandplimittool/).
>
> Hope this helps.
>
> Regards,
> Sripad
>
>
> -----Original Message-----
> From: Confocal Microscopy List [mailto:[hidden email]]
> On
> Behalf Of Christophe Leterrier
> Sent: Tuesday, March 04, 2014 3:16 PM
> To: [hidden email]
> Subject: Re: discussion is open...continues ... about localization and
> super
> resolution
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> This is a great introduction to a question a wanted to ask to the SMLM
> specialists (I did not find the answer in the above reference). I also
> posted it on the Nanoscopy LinkedIn group, so I'm sorry for people that
> will
> get it two times!
>
> There are several formulas to estimate the lateral (XY) precision of
> individual localization events in PALM/STORM microscopy. This goes from the
> crude sigma(PSF)/√Nphotons to more elaborate formulas that includes
> pixelation and background (such as Thomson & Webb Biophys J 2002, corrected
> in Mortensen Nat Methods 2010). What I'm looking for is an equation that
> allows to estimate the precision for the Z localization when using
> astigmatism-based 3D PALM/STORM. The reference I have found where this
> precision is studied theoretically (such as Badierostami Applied Phys Lett
> 2010, Shaevitz Int J Optics 2009) do not use a directly calculable formula.
>
> I assume the precision depends on the photon number and the point on the Z
> calibration curve with the respective X and Y PSF width. Is anyone aware of
> such a formula?
>
> Best Regards,
>
> Christophe
>
>
>
>
> 2014-03-04 21:52 GMT+01:00 Alberto Diaspro <[hidden email]>:
>
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > *****
> >
> > http://www.nature.com/nmeth/journal/v11/n3/full/nmeth.2843.html
> >"
>