Re: Quantifying fluorescence help
Posted by George McNamara on
URL: http://confocal-microscopy-list.275.s1.nabble.com/Quantifying-fluorescence-help-tp593618p593620.html
Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hi Maria,
See current issue of JCB:
Quantitative analysis of autophagy-related protein stoichiometry by fluorescence microscopy
Some suggestions from me:
* as noted by others, read Pawley's "39" article in Biotechiques
* use as low magnification (but decent numerical aperture, for good light throughput) lens. low mag => more cells per field of view. More cells => higher N. "N is good".
Decent NA means fewer optical sections to obtain complete Z-series (ideal might be one optical slice).
* Verify the instrument is producing stable results. Can be done by fluorescent beads. consider including some beads 9of comparable brightness to your cells) in with the samples. See very bottom of this message for a DIY test (or better yet, have the confocal core manager do it, so it is their dime, not yours - get the results from several nights in a row). If the lasers are not stable, your results will not be.
* consider centrifuging (lightly) the cells onto the slide, and/or other methods, so they are all in the same orientation (not some lying down, others nose/tail up - I'm much better with my coral fluorescent protein anatomy and taxonomy than for zooxanthellae: do dino's have noses?).
* these are fixed cells, right? If so, refractive index match the mounting and immersion medium, i.e. 25x/0.8 NA IMM lens with immersion oil and 2,2'-thiodiethanol (both RI 1.515, I consider 25x lens a moderate mag). With respect to TDE, read the Staudt, Hell et al 2007 Microsc Res Tech article and take the time to understand what the "ocean of fluorescence" vs RI and depth graph means for your study (hint: a perfect confocal microscope and specimen would produce a horizontal line at 1.0 relative brightness).
* if using multiple wavelengths, adjust the pinhole size(s) so that the same optical slice size is being acquired at every fluorescent wavelength. For example, if "red' 1.0 Airy unit is a 1.0 um optical slice size, set the "green" channel optical slice size to 1.0 um as well.
* You get for free the shape of the cell (and all the membranes inside it) by adding a reflected light channel. This may be helpful for segmentation (or maybe not). Reflection mode also lets you find the coverglass (very bright!) and slide (bright) surfaces. You can use 1.0 Airy units for reflection mode, as long as you are not trying to "colocalize" with the "matched thickness" fluorescence channels.
* I usually acquire Z-series with a physical step size either (a) 1/2 optical slice thickness ("overlap mode"), if bright specimen with minimal photobleaching, or (b) equal to optical slice thickness (if bleaching or on my dime). Nyquist purists would argue for 1/2.3 or even 1/3 of optical slice thickness, but then Nyquist did not have specimens that photobleached.
As Pawley wrote, "deconvolve everything" (to which I'll add; with the right algorithm, with a well functioning instrument, and verify that the output makes sense).
More interesting references:
Kai K, Kitajima Y, Hiraki M, Satoh S, Tanaka M, Nakafusa Y, Tokunaga O, Miyazaki K.
Quantitative double-fluorescence immunohistochemistry (qDFIHC), a
novel technology to assess protein expression: a pilot study analyzing
5-FU sensitive markers thymidylate synthase, dihydropyrimidine
dehydrogenase and orotate phosphoribosyl transferases in gastric cancer
tissue specimens. Cancer Lett. 2007 Dec 8;258(1):45-54. PMID:
17892912 (in my opinion: not perfect [what is] but not
bad).
Swedlow JR, Hu K, Andrews PD, Roos DS, Murray JM.
Measuring tubulin content in Toxoplasma gondii: a comparison of laser-scanning confocal and wide-field fluorescence microscopy.
Proc Natl Acad Sci U S A. 2002 Feb 19;99(4):2014-9. PMID: 11830634 (experienced microscopists, but how hard did they try to get confocal to work well? also, I do not recall them deconvolving [with a confocal optimized algorithm] the confocal data).
best wishes,
George
At 09:25 AM 7/18/2008, you wrote:
George McNamara, Ph.D.
University of Miami, Miller School of Medicine
Image Core
Miami, FL 33010
[hidden email]
[hidden email]
305-243-8436 office
http://home.earthlink.net/~pubspectra/
http://home.earthlink.net/~geomcnamara/
http://www.sylvester.org/research/SR_lab_analytical.asp?ana=desc (Analytical Imaging Core Facility)
simple laser performance test to propose:
Zeiss LSM510 jargon (ChD is transmitted light detector)
4-tracks:
633 nm -> ChD
543 nm -> ChD
488 nm -> ChD
364 nm -> ChD
(laser wavelengths can be substituted, but these are the four major channels used for immunofluorescence, so makes sense to test these. The old "LSM" software is limited to 4 tracks, 8 channels).
Use same gain setting for all ChD tracks (ex. 260 gain, 0 offset, making sure no under or over saturation). Adjust laser power to get on scale (aim for between 2000-3000 gray level to start with). Recommendation: 1.0 or 2.0 OD filter above condenser (for an inverted microscope). I leave my 1.0 OD filter taped down on the condenser dovetail, just below where the condenser polarizer swings in place.
512x512 pixel mode, 12-bit (I forget what scan speed I use, it is probably the "default").
No specimen in place (since having one could result in focus drift leading to change in brightness)
10x/0.3 NA lens is what I use. Adjust condenser focus so condenser field aperture diaphragm (CFAD) is in focus (i.e. Kohler illumination, albeit, no specimen). Adjust CFAD so it is centered and spans ~90% of the field of view (this way the corners give you the 'dark current").
Turn on the entire system (i.e. monday AM), get all this configured (i.e. "Reuse" button from a previous session), do timelapse at 1 minute interval (until first user of the week). After last user is done, run overnight. If you are (un)lucky, no user for first day, which is the best way to see how long the lasers take to reach stable operating condition. If the facility is busy, turn off the lasers for 1 hour Friday afternoon, turn on, start immediately, run over the weekend (warning: on my LSM system, it often has problems correctly saving 2000+ timepoint series). At end (or at a break between users) I place a large circular ROI in the illuminated area (avoiding the dimmer ring of the CFAD edge), save the
Expected results:
1. good: "rock solid" (less than 10% change from min to max, and no sudden step changes) if the lasers have been on for at least 24 hours. I run mine Monday AM to Friday night (or for multiple weeks, if users have told me they might use the system over the weekend and heavy rains [hurricane] not expected). I do have a service contract with Zeiss.
2. bad: if your instrument overheats (ex. no one has ever removed the dust from the bottom of the real time computer of the LSM510 [the solution to major instabilities during the first several months of my managing this unit, until field service engineer #3 did a much more thorough job and de-dusted the RTC], and/or you have a crappy AOTF), you will see ugly changes in power in any or all wavelengths.
Note: my 364 nm line takes about 8-24 hours to reach "stable" operating condition. It always produces ugly lines in the ChD test (and DIC is ugly). Fluorescence of DAPI looks fine.
Some folks at Zeiss used to suggest that ChD detector do not report comparable intensities to the fluorescence/reflection PMTs. I have performed similar test through a fluorescent Chroma plastic slide (focused in plastic and relatively open pinhole to avoid potential focus drift "issue"), the ChD and Ch# channels track perfectly.
URL: http://confocal-microscopy-list.275.s1.nabble.com/Quantifying-fluorescence-help-tp593618p593620.html
Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Hi Maria,
See current issue of JCB:
Quantitative analysis of autophagy-related protein stoichiometry by fluorescence microscopy
- Jiefei Geng, Misuzu Baba, Usha Nair, and Daniel J. Klionsky
- J. Cell Biol. 2008 182: 129-140. Published Jul 14 2008, 10.1083/jcb.200711112. [Abstract] [Full Text] [PDF] [Supplemental Material index]
Some suggestions from me:
* as noted by others, read Pawley's "39" article in Biotechiques
* use as low magnification (but decent numerical aperture, for good light throughput) lens. low mag => more cells per field of view. More cells => higher N. "N is good".
Decent NA means fewer optical sections to obtain complete Z-series (ideal might be one optical slice).
* Verify the instrument is producing stable results. Can be done by fluorescent beads. consider including some beads 9of comparable brightness to your cells) in with the samples. See very bottom of this message for a DIY test (or better yet, have the confocal core manager do it, so it is their dime, not yours - get the results from several nights in a row). If the lasers are not stable, your results will not be.
* consider centrifuging (lightly) the cells onto the slide, and/or other methods, so they are all in the same orientation (not some lying down, others nose/tail up - I'm much better with my coral fluorescent protein anatomy and taxonomy than for zooxanthellae: do dino's have noses?).
* these are fixed cells, right? If so, refractive index match the mounting and immersion medium, i.e. 25x/0.8 NA IMM lens with immersion oil and 2,2'-thiodiethanol (both RI 1.515, I consider 25x lens a moderate mag). With respect to TDE, read the Staudt, Hell et al 2007 Microsc Res Tech article and take the time to understand what the "ocean of fluorescence" vs RI and depth graph means for your study (hint: a perfect confocal microscope and specimen would produce a horizontal line at 1.0 relative brightness).
* if using multiple wavelengths, adjust the pinhole size(s) so that the same optical slice size is being acquired at every fluorescent wavelength. For example, if "red' 1.0 Airy unit is a 1.0 um optical slice size, set the "green" channel optical slice size to 1.0 um as well.
* You get for free the shape of the cell (and all the membranes inside it) by adding a reflected light channel. This may be helpful for segmentation (or maybe not). Reflection mode also lets you find the coverglass (very bright!) and slide (bright) surfaces. You can use 1.0 Airy units for reflection mode, as long as you are not trying to "colocalize" with the "matched thickness" fluorescence channels.
* I usually acquire Z-series with a physical step size either (a) 1/2 optical slice thickness ("overlap mode"), if bright specimen with minimal photobleaching, or (b) equal to optical slice thickness (if bleaching or on my dime). Nyquist purists would argue for 1/2.3 or even 1/3 of optical slice thickness, but then Nyquist did not have specimens that photobleached.
As Pawley wrote, "deconvolve everything" (to which I'll add; with the right algorithm, with a well functioning instrument, and verify that the output makes sense).
More interesting references:
Kai K, Kitajima Y, Hiraki M, Satoh S, Tanaka M, Nakafusa Y, Tokunaga O, Miyazaki K.
Quantitative double-fluorescence immunohistochemistry (qDFIHC), a
novel technology to assess protein expression: a pilot study analyzing
5-FU sensitive markers thymidylate synthase, dihydropyrimidine
dehydrogenase and orotate phosphoribosyl transferases in gastric cancer
tissue specimens. Cancer Lett. 2007 Dec 8;258(1):45-54. PMID:
17892912 (in my opinion: not perfect [what is] but not
bad).Swedlow JR, Hu K, Andrews PD, Roos DS, Murray JM.
Measuring tubulin content in Toxoplasma gondii: a comparison of laser-scanning confocal and wide-field fluorescence microscopy.
Proc Natl Acad Sci U S A. 2002 Feb 19;99(4):2014-9. PMID: 11830634 (experienced microscopists, but how hard did they try to get confocal to work well? also, I do not recall them deconvolving [with a confocal optimized algorithm] the confocal data).
best wishes,
George
At 09:25 AM 7/18/2008, you wrote:
Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
I am a PhD student at Auburn University just getting started into my thesis work. I am
looking for a reliable method for quantifying fluorescence.
My project deals with marine dinoflagellates (zooxanthellae) that reside intracellularly in
hosts (usually cnidarians). I am working with cultures or isolates of only the dinoflagellates
for my confocal work. I have an antibody that was created against the surface secretions
of mucilage (secreted as part of a daily cycle by the alga) from one strain of zooxanthellae.
I am attempting to use this antibody to label various strains to identify differences in
mucilage between them. Thus far, I have seen that the strain the antibody was created
against labels around the cell fairly brightly. Most samples either show this or a complete
lack of labeling. However, a few samples show a faint fluorescence lifted off the cell
surface. I would like to be able to quantify this fluorescence in comparison to either the
control strain that the antibody was created against or against a known fluorescence.
Any help on ideas for this, or places to look for ideas would be greatly appreciated.
Thanks!
Maria Mazzillo
George McNamara, Ph.D.
University of Miami, Miller School of Medicine
Image Core
Miami, FL 33010
[hidden email]
[hidden email]
305-243-8436 office
http://home.earthlink.net/~pubspectra/
http://home.earthlink.net/~geomcnamara/
http://www.sylvester.org/research/SR_lab_analytical.asp?ana=desc (Analytical Imaging Core Facility)
simple laser performance test to propose:
Zeiss LSM510 jargon (ChD is transmitted light detector)
4-tracks:
633 nm -> ChD
543 nm -> ChD
488 nm -> ChD
364 nm -> ChD
(laser wavelengths can be substituted, but these are the four major channels used for immunofluorescence, so makes sense to test these. The old "LSM" software is limited to 4 tracks, 8 channels).
Use same gain setting for all ChD tracks (ex. 260 gain, 0 offset, making sure no under or over saturation). Adjust laser power to get on scale (aim for between 2000-3000 gray level to start with). Recommendation: 1.0 or 2.0 OD filter above condenser (for an inverted microscope). I leave my 1.0 OD filter taped down on the condenser dovetail, just below where the condenser polarizer swings in place.
512x512 pixel mode, 12-bit (I forget what scan speed I use, it is probably the "default").
No specimen in place (since having one could result in focus drift leading to change in brightness)
10x/0.3 NA lens is what I use. Adjust condenser focus so condenser field aperture diaphragm (CFAD) is in focus (i.e. Kohler illumination, albeit, no specimen). Adjust CFAD so it is centered and spans ~90% of the field of view (this way the corners give you the 'dark current").
Turn on the entire system (i.e. monday AM), get all this configured (i.e. "Reuse" button from a previous session), do timelapse at 1 minute interval (until first user of the week). After last user is done, run overnight. If you are (un)lucky, no user for first day, which is the best way to see how long the lasers take to reach stable operating condition. If the facility is busy, turn off the lasers for 1 hour Friday afternoon, turn on, start immediately, run over the weekend (warning: on my LSM system, it often has problems correctly saving 2000+ timepoint series). At end (or at a break between users) I place a large circular ROI in the illuminated area (avoiding the dimmer ring of the CFAD edge), save the
Expected results:
1. good: "rock solid" (less than 10% change from min to max, and no sudden step changes) if the lasers have been on for at least 24 hours. I run mine Monday AM to Friday night (or for multiple weeks, if users have told me they might use the system over the weekend and heavy rains [hurricane] not expected). I do have a service contract with Zeiss.
2. bad: if your instrument overheats (ex. no one has ever removed the dust from the bottom of the real time computer of the LSM510 [the solution to major instabilities during the first several months of my managing this unit, until field service engineer #3 did a much more thorough job and de-dusted the RTC], and/or you have a crappy AOTF), you will see ugly changes in power in any or all wavelengths.
Note: my 364 nm line takes about 8-24 hours to reach "stable" operating condition. It always produces ugly lines in the ChD test (and DIC is ugly). Fluorescence of DAPI looks fine.
Some folks at Zeiss used to suggest that ChD detector do not report comparable intensities to the fluorescence/reflection PMTs. I have performed similar test through a fluorescent Chroma plastic slide (focused in plastic and relatively open pinhole to avoid potential focus drift "issue"), the ChD and Ch# channels track perfectly.
| Free forum by Nabble | Edit this page |