Poor man's confocal
Locked
|
Anybody have any experience with the Zeiss Apotome structured illumination system?
I am writing a proposal for a new microscope system and the vendor mentioned this as a way
to reduce background in fluorescence microscopy.
Apparently, it superimposes a moving grid on the image, and this somehow allows background
noise reduction.
We have used deconvolution to do this, but find that it takes a long time to do.
I think it would add about $20k to the cost of the scope.
Opinions? Worthwhile? Any experience with it?
Bob Nienhuis
Neurobiology Research M/S151A3
UCLA / VA Medical Center
North Hills, CA
|
|
|
There has been a great deal of discusion of structured illumination
recently. You might take a look at NATUREMETHODS VOL.6NO.5:339 for one example. In addition to the Apotome modification, there is also an add-on for other microscopes by Optigrid. (http://www.qioptiqlinos.com/Products/StructuredLightSystem/) I haven't used either system, but I am also interested. Let me know what you find out. Joel -------------- Original message --------------- Anybody have any experience with the Zeiss Apotome structured illumination system? I am writing a proposal for a new microscope system and the vendor mentioned this as a way to reduce background in fluorescence microscopy. Apparently, it superimposes a moving grid on the image, and this somehow allows background noise reduction. We have used deconvolution to do this, but find that it takes a long time to do. I think it would add about $20k to the cost of the scope. Opinions? Worthwhile? Any experience with it? Bob Nienhuis Neurobiology Research M/S151A3 UCLA / VA Medical Center North Hills, CA [hidden email] -- Joel B. Sheffield, Ph.D. Biology Department, Temple University 1900 North 12th Street Philadelphia, PA 19122 [hidden email] (215) 204 8839, fax (215) 204 0486 http://astro.temple.edu/~jbs |
|
|
I have just finished training a user who came to us because the Apotome
system did not give anything like the results of the LSM510. I certainly don't mean to suggest the Apotome system doesn't work, but you should get a demo on your samples to see how it works for you. Dale Joel Sheffield wrote: > There has been a great deal of discusion of structured illumination > recently. You might take a look at NATUREMETHODS VOL.6NO.5:339 > for one example. In addition to the Apotome modification, there is > also an add-on for other microscopes by Optigrid. > (http://www.qioptiqlinos.com/Products/StructuredLightSystem/) > > I haven't used either system, but I am also interested. Let me know > what you find out. > > Joel > > -------------- Original message --------------- > Anybody have any experience with the Zeiss Apotome structured > illumination system? > > I am writing a proposal for a new microscope system and the vendor > mentioned this as a way > to reduce background in fluorescence microscopy. > > Apparently, it superimposes a moving grid on the image, and this > somehow allows background > noise reduction. > > We have used deconvolution to do this, but find that it takes a long > time to do. > > I think it would add about $20k to the cost of the scope. > > Opinions? Worthwhile? Any experience with it? > > Bob Nienhuis > Neurobiology Research M/S151A3 > UCLA / VA Medical Center > North Hills, CA > [hidden email] > -- > Joel B. Sheffield, Ph.D. > Biology Department, Temple University > 1900 North 12th Street > Philadelphia, PA 19122 > [hidden email] > (215) 204 8839, fax (215) 204 0486 > http://astro.temple.edu/~jbs |
|
|
I been told the images of the apotome are not quantitative... If that is true, that seems to me a big limitation of using the apotome for any serious fluorescence microscopy application.
Leoncio -----Original Message----- From: Confocal Microscopy List [mailto:[hidden email]] On Behalf Of Dale Callaham Sent: Wednesday, May 26, 2010 2:52 PM To: [hidden email] Subject: Re: Poor man's confocal I have just finished training a user who came to us because the Apotome system did not give anything like the results of the LSM510. I certainly don't mean to suggest the Apotome system doesn't work, but you should get a demo on your samples to see how it works for you. Dale Joel Sheffield wrote: > There has been a great deal of discusion of structured illumination > recently. You might take a look at NATUREMETHODS VOL.6NO.5:339 for > one example. In addition to the Apotome modification, there is also > an add-on for other microscopes by Optigrid. > (http://www.qioptiqlinos.com/Products/StructuredLightSystem/) > > I haven't used either system, but I am also interested. Let me know > what you find out. > > Joel > > -------------- Original message --------------- Anybody have any > experience with the Zeiss Apotome structured illumination system? > > I am writing a proposal for a new microscope system and the vendor > mentioned this as a way to reduce background in fluorescence > microscopy. > > Apparently, it superimposes a moving grid on the image, and this > somehow allows background noise reduction. > > We have used deconvolution to do this, but find that it takes a long > time to do. > > I think it would add about $20k to the cost of the scope. > > Opinions? Worthwhile? Any experience with it? > > Bob Nienhuis > Neurobiology Research M/S151A3 > UCLA / VA Medical Center > North Hills, CA > [hidden email] > -- > Joel B. Sheffield, Ph.D. > Biology Department, Temple University > 1900 North 12th Street > Philadelphia, PA 19122 > [hidden email] > (215) 204 8839, fax (215) 204 0486 > http://astro.temple.edu/~jbs |
|
|
In reply to this post by Bob Nienhuis
Hi Bob, The performance of a structured illumination system may be close to the capabilities of a confocal microscope but this cannot be an alternative for CLSM. The performance difference is considerably high and Apotome has no flexibility like confocal. This can be used as entry level optical sectioning equipment. In my opinion, this can be used as a pre-scanning system before we do specimen analysis with CLSM. We can filter the number of specimen to be analyzed with CLSM. If I am not mistaken, Apotome cannot be fitted with any other microscope as it is designed for Zeiss systems. However, OptiGrid can be fitted with any existing microscopes. In addition, the buyer can select the CCD/EMCCD of their choice. Why not consider going for an entry level CLSM like Nikon C1 plus or Olympus FV10i that may cost less than or almost the same cost of Apotome + Microscope, but you get the real uncompromising result of confocal microscopy? Roshma. On Thu, May 27, 2010 at 12:02 AM, Bob Nienhuis <[hidden email]> wrote:
|
|
|
I've found that for thin, fixed specimen (i.e. cultured cells), the apotome images quality approaches that of confocal images. However I was surprised how high the photobleaching was, it was significantly higher than our entry level Leica TCS-SPE confocal. Maybe with a more sensitive camera (we have the regular Zeiss one) this could be improved.
Christophe -- Christophe Leterrier Postdoc INSERM UMR641 Neurobiologie des Canaux Ioniques IFR Jean Roche, Université de la Méditerranée Marseille, France [hidden email] On Thu, May 27, 2010 at 06:05, Roshma Azeem <[hidden email]> wrote:
|
|
|
In reply to this post by Roshma Azeem
Well, Dear Confocalists.
What will you get with the entry level CLSM?
It is very for students to learn microscopy, no doubt!
Also you can perform rather limited physiologically relevant research with abundant biomolecules, e.g. actin, tubulins, histons. The images of cytoskeleton are truly magnificent - many of them certainly deserve to be placed in the Art Gallery in Paris - but not anymore in the Cell, Science or Nature magazines, unless ....
Broadly speaking, all other applications require >>100-fold overexpression of a molecule of interest. At these concentrations biomolecules tend to stick to each rather than functionally interact (Kd of EGFP, for example is ca. 100-200uM, when tagged to another protein this value often goes down a lot). That is the main reason why the high end, physiologically relevant research is very slow and unproductive.
Another tiny peculiarity: none of the CLSM systems offers an optimal (i.e. compromise-free), zero-cross talk three channel imaging of, let say, CeFP, EYFP, mRFP1/mCherry tagged biomolecules that allow highly accurate quantitative image analysis on a voxel-by-voxel basis.
Dear CLSM Senior salesmen and VPs of Marketing - it is already 2010, ten years into the New Millennium.
Welcome to a hopefully fruitful discussion on this highly relevant topic.
I give you a very typical example: in 2006 Coherent demoed 570 and 580nm laser lines. Thanks to the "Powerful Marketing", three years later, (only three years!, "so quick!"), it ended up with the 568 nm sapphire (instead of the 572 nm line), and with the 1000 mW 577 nm line, which again, thanks to a "careful marketing study", has been designed for the clinical medicine rather than for the advanced biomedical research. I do hope, that still in this Millennium, the working inter-industrial, academio-industrial bridges will be built between the Senior VPs of Marketing and advanced sciences - at this point there are plenty of gaps to be filled. And at the current pace of interdisciplinary integration it sounds it will take a century or two to succeed.
Best,
Vitaly
301-515-7833
From: Roshma Azeem <[hidden email]> To: [hidden email] Sent: Thu, May 27, 2010 6:05:43 AM Subject: Re: Poor man's confocal Hi Bob, The performance of a structured illumination system may be close to the capabilities of a confocal microscope but this cannot be an alternative for CLSM. The performance difference is considerably high and Apotome has no flexibility like confocal. This can be used as entry level optical sectioning equipment. In my opinion, this can be used as a pre-scanning system before we do specimen analysis with CLSM. We can filter the number of specimen to be analyzed with CLSM. If I am not mistaken, Apotome cannot be fitted with any other microscope as it is designed for Zeiss systems. However, OptiGrid can be fitted with any existing microscopes. In addition, the buyer can select the CCD/EMCCD of their choice. Why not consider going for an entry level CLSM like Nikon C1 plus or Olympus FV10i that may cost less than or almost the same cost of Apotome + Microscope, but you get the real uncompromising result of confocal microscopy? Roshma. On Thu, May 27, 2010 at 12:02 AM, Bob Nienhuis <[hidden email]> wrote:
|
|
|
In reply to this post by Bob Nienhuis
Hi Bob,
On May 27, 2010, at 7:02 AM, CONFOCALMICROSCOPY automatic digest system wrote: > Date: Wed, 26 May 2010 11:32:12 -0700 > From: Bob Nienhuis <[hidden email]> > Subject: Poor man's confocal > > Anybody have any experience with the Zeiss Apotome structured illumination > system? Yes, I works and does what its supposed to do pretty well. However there are limitations, and one should not attempt to compare it closely with a point scanning confocal, its not supposed to replace a confocal, rather its a very different approach, that happens to be relatively cheap. Apotome and Optigrid are basically rather similar.... the limitations are: 1) that you must take several (eg 3) images of each filed of view - so bleaching can become a problem 2) if the lamp is unstable then some fo the 3 images are darker than others, this messed up the final image. 3) One must carefully align the grid in the in its position conjugate to the image plane, since you need to hgrid parrern to have as good contrast as you can get for the mthod ot work. 4) very thick samples or samples with very few features will not work. 5) the result can be a bit noisy (compared to the input images) , because the simple algorithm used throw away much of the signal. 6) Dont try to measure relative fluorescence intensities ... the method is not linear. good things are: 1) Price, its simple and cheap 2) Speed - pretty fast compared to point scanning a large field of vire 3) no other special stuff needed - same stand and filters and objectives etc. 4) Great for screening samples / yes or no answers. > > I am writing a proposal for a new microscope system and the vendor mentioned > this as a way > to reduce background in fluorescence microscopy. It works, but there is, also quite cheap technology coming out now, which i think does a better job: Zeiss Viva Tome (3 colours) and the same gadget in a different box from Andor (4 colours possible apparently) Method is called aperture correlation, and its like a mixture of spinning disk confocal and structured illumination with collection of out of focus and mostly infocus signal, then an attempt to remove most of the residual blurr from the moistly in focus signal my subtracting a weighted out of focus image. It works really well (I saw it live at ELMI 2010) and is faster then apotom/optigrid as only one CCD exposure is needed. So less bleaching. I want one. I would say aperture correlation "seems better" than the older simpestructured illumination methods. Super resolution structured illumination methods like the Zeiss Elyra and Applies precision OMX are a totally different story... dont confuse them with the apo tome/via tome stuff. > > Apparently, it superimposes a moving grid on the image, and this somehow > allows background > noise reduction. look here for apotome http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/apotomemechanics/index.html and here for vivatome http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/vivatomeopticaltrain/index.html aperture correlation method comes from Tony wilson's lab, see more references here http://zeiss-campus.magnet.fsu.edu/referencelibrary/aperturecorrelation.html > > We have used deconvolution to do this, but find that it takes a long time to > do. its doenst have to take that long if you get a good feel for ot and have a computer with tens of processors. > > I think it would add about $20k to the cost of the scope. > > Opinions? apotome - good for what its designed for - but its not a poor mans confocal, its something rather different. > Worthwhile? definately > Any experience with it? yes, it works great for thin bright fixed samples. but vivatome looks better. cheers Dan > > Bob Nienhuis > Neurobiology Research M/S151A3 > UCLA / VA Medical Center > North Hills, CA > [hidden email] Dr. Daniel James White BSc. (Hons.) PhD Senior Microscopist / Image Visualisation, Processing and Analysis Light Microscopy and Image Processing Facilities Max Planck Institute of Molecular Cell Biology and Genetics Pfotenhauerstrasse 108 01307 DRESDEN Germany +49 (0)15114966933 (German Mobile) +49 (0)351 210 2627 (Work phone at MPI-CBG) +49 (0)351 210 1078 (Fax MPI-CBG LMF) http://www.bioimagexd.net BioImageXD http://pacific.mpi-cbg.de Fiji - is just ImageJ (Batteries Included) http://www.chalkie.org.uk Dan's Homepages https://ifn.mpi-cbg.de Dresden Imaging Facility Network dan (at) chalkie.org.uk ( white (at) mpi-cbg.de ) |
|
|
In reply to this post by Vitaly Boyko
Vitaly, live cell imaging of transfected fusion proteins is an important aspect, but is not the only application of CLSM... I usually use widefield for live cell imaging, to have a better sensitivity, and CLSM for immunocytochemistry.
Christophe On Thu, May 27, 2010 at 10:54, Vitaly Boyko <[hidden email]> wrote:
|
|
|
In reply to this post by Daniel James White
There is at least another cheap and smart method to get optical
sectioning from a widefield microscope, called "HiLo" microscopy: Wide-field fluorescence sectioning with hybrid speckle and uniform-illumination microscopy. Lim D, Chu KK, Mertz J. Opt Lett. 2008 Aug 15;33(16):1819-21. The idea is to take a speckled image and a non speckled one, from which high and low frequencies of the images are computed and merged to from the final image. I don't think there is a commercial version of it, but I've seen two other labs using it and it seems quite simple to implement. Christophe On Thu, May 27, 2010 at 11:03, Daniel James White <[hidden email]> wrote: > > Hi Bob, > > > On May 27, 2010, at 7:02 AM, CONFOCALMICROSCOPY automatic digest system wrote: > > > Date: Wed, 26 May 2010 11:32:12 -0700 > > From: Bob Nienhuis <[hidden email]> > > Subject: Poor man's confocal > > > > Anybody have any experience with the Zeiss Apotome structured illumination > > system? > > Yes, I works and does what its supposed to do pretty well. > However there are limitations, and one should not attempt to compare it closely with a point scanning confocal, > its not supposed to replace a confocal, rather its a very different approach, that happens to be relatively cheap. > > Apotome and Optigrid are basically rather similar.... > the limitations are: > 1) that you must take several (eg 3) images of each filed of view - so bleaching can become a problem > 2) if the lamp is unstable then some fo the 3 images are darker than others, this messed up the final image. > 3) One must carefully align the grid in the in its position conjugate to the image plane, > since you need to hgrid parrern to have as good contrast as you can get for the mthod ot work. > 4) very thick samples or samples with very few features will not work. > 5) the result can be a bit noisy (compared to the input images) , because the simple algorithm used throw away much of the signal. > 6) Dont try to measure relative fluorescence intensities ... the method is not linear. > good things are: > 1) Price, its simple and cheap > 2) Speed - pretty fast compared to point scanning a large field of vire > 3) no other special stuff needed - same stand and filters and objectives etc. > 4) Great for screening samples / yes or no answers. > > > > > > I am writing a proposal for a new microscope system and the vendor mentioned > > this as a way > > to reduce background in fluorescence microscopy. > > It works, > but there is, also quite cheap technology coming out now, which i think does a better job: > > Zeiss Viva Tome (3 colours) and the same gadget in a different box from Andor (4 colours possible apparently) > Method is called aperture correlation, and its like a mixture of spinning disk confocal and structured illumination > with collection of out of focus and mostly infocus signal, then an attempt to remove most of the residual blurr from the > moistly in focus signal my subtracting a weighted out of focus image. > > It works really well (I saw it live at ELMI 2010) and is faster then apotom/optigrid as only one CCD exposure is needed. > So less bleaching. > > I want one. I would say aperture correlation "seems better" than the older simpestructured illumination methods. > > Super resolution structured illumination methods like the Zeiss Elyra and Applies precision OMX > are a totally different story... dont confuse them with the apo tome/via tome stuff. > > > > > Apparently, it superimposes a moving grid on the image, and this somehow > > allows background > > noise reduction. > > look here for apotome > http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/apotomemechanics/index.html > and here for vivatome > http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/vivatomeopticaltrain/index.html > > aperture correlation method comes from Tony wilson's lab, see more references here > http://zeiss-campus.magnet.fsu.edu/referencelibrary/aperturecorrelation.html > > > > > We have used deconvolution to do this, but find that it takes a long time to > > do. > > its doenst have to take that long if you get a good feel for ot and have a computer with tens of processors. > > > > > > I think it would add about $20k to the cost of the scope. > > > > Opinions? > > apotome - good for what its designed for - but its not a poor mans confocal, its something rather different. > > > Worthwhile? > > definately > > > Any experience with it? > > yes, it works great for thin bright fixed samples. > > but vivatome looks better. > > > cheers > > Dan > > > > > > > Bob Nienhuis > > Neurobiology Research M/S151A3 > > UCLA / VA Medical Center > > North Hills, CA > > [hidden email] > > Dr. Daniel James White BSc. (Hons.) PhD > Senior Microscopist / Image Visualisation, Processing and Analysis > Light Microscopy and Image Processing Facilities > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > > http://www.bioimagexd.net BioImageXD > http://pacific.mpi-cbg.de Fiji - is just ImageJ (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Dresden Imaging Facility Network > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) |
|
|
In reply to this post by Bob Nienhuis
Bob,
I've used the Zeiss Apotome to image neuromuscular junctions labeled with three fluorophores in muscle tissue 50 microns thick. The images weren't as pretty as confocal images, but I could resolve my structures of interest and it provided me with a nice 3D image from which I could make morphological measurements. Whether or not it would work for your application all depends on what you're imaging and what you're measuring. If you're currenlty using widefield and deconvolution to achieve your final images, I imagine the Apotome might be a good replacement. I'm sure the folks at Zeiss would be happy to come demo a system, and that would be your best test. Aaron --------------- Aaron Johnson Pre-Doctoral Fellow Department of Surgery University of Wisconsin, Madison |
|
|
In reply to this post by Daniel James White
Good summary. It depends on how brightly your samples are labeled and your resolution needs.
But, for the additional US $20,000, you can purchase a lot of computing power to speed the deconvolution and gain a more flexible tool. That power can be applied to images collected by other modalities. It can also be put to work for analysis since deconvolution can produce quantitative images. You may want to investigate means by which decon may be speeded up, as in a different algorithm or implementation, improve the SNR of your images through optimizing the specimen (are specimen optics matched to the imaging system) and perhaps reducing the number of voxels that need to be deconvolved. eg. A full set of iterations may not be required for every channel. Regards, Glen Glen MacDonald Core for Communication Research Virginia Merrill Bloedel Hearing Research Center Box 357923 University of Washington Seattle, WA 98195-7923 USA (206) 616-4156 [hidden email] On May 27, 2010, at 2:03 AM, Daniel James White wrote: > Hi Bob, > > > On May 27, 2010, at 7:02 AM, CONFOCALMICROSCOPY automatic digest system wrote: > >> Date: Wed, 26 May 2010 11:32:12 -0700 >> From: Bob Nienhuis <[hidden email]> >> Subject: Poor man's confocal >> >> Anybody have any experience with the Zeiss Apotome structured illumination >> system? > > Yes, I works and does what its supposed to do pretty well. > However there are limitations, and one should not attempt to compare it closely with a point scanning confocal, > its not supposed to replace a confocal, rather its a very different approach, that happens to be relatively cheap. > > Apotome and Optigrid are basically rather similar.... > the limitations are: > 1) that you must take several (eg 3) images of each filed of view - so bleaching can become a problem > 2) if the lamp is unstable then some fo the 3 images are darker than others, this messed up the final image. > 3) One must carefully align the grid in the in its position conjugate to the image plane, > since you need to hgrid parrern to have as good contrast as you can get for the mthod ot work. > 4) very thick samples or samples with very few features will not work. > 5) the result can be a bit noisy (compared to the input images) , because the simple algorithm used throw away much of the signal. > 6) Dont try to measure relative fluorescence intensities ... the method is not linear. > good things are: > 1) Price, its simple and cheap > 2) Speed - pretty fast compared to point scanning a large field of vire > 3) no other special stuff needed - same stand and filters and objectives etc. > 4) Great for screening samples / yes or no answers. > > >> >> I am writing a proposal for a new microscope system and the vendor mentioned >> this as a way >> to reduce background in fluorescence microscopy. > > It works, > but there is, also quite cheap technology coming out now, which i think does a better job: > > Zeiss Viva Tome (3 colours) and the same gadget in a different box from Andor (4 colours possible apparently) > Method is called aperture correlation, and its like a mixture of spinning disk confocal and structured illumination > with collection of out of focus and mostly infocus signal, then an attempt to remove most of the residual blurr from the > moistly in focus signal my subtracting a weighted out of focus image. > > It works really well (I saw it live at ELMI 2010) and is faster then apotom/optigrid as only one CCD exposure is needed. > So less bleaching. > > I want one. I would say aperture correlation "seems better" than the older simpestructured illumination methods. > > Super resolution structured illumination methods like the Zeiss Elyra and Applies precision OMX > are a totally different story... dont confuse them with the apo tome/via tome stuff. > >> >> Apparently, it superimposes a moving grid on the image, and this somehow >> allows background >> noise reduction. > > look here for apotome > http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/apotomemechanics/index.html > and here for vivatome > http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/vivatomeopticaltrain/index.html > > aperture correlation method comes from Tony wilson's lab, see more references here > http://zeiss-campus.magnet.fsu.edu/referencelibrary/aperturecorrelation.html > >> >> We have used deconvolution to do this, but find that it takes a long time to >> do. > > its doenst have to take that long if you get a good feel for ot and have a computer with tens of processors. > > >> >> I think it would add about $20k to the cost of the scope. >> >> Opinions? > > apotome - good for what its designed for - but its not a poor mans confocal, its something rather different. > >> Worthwhile? > > definately > >> Any experience with it? > > yes, it works great for thin bright fixed samples. > > but vivatome looks better. > > > cheers > > Dan > > > >> >> Bob Nienhuis >> Neurobiology Research M/S151A3 >> UCLA / VA Medical Center >> North Hills, CA >> [hidden email] > > Dr. Daniel James White BSc. (Hons.) PhD > Senior Microscopist / Image Visualisation, Processing and Analysis > Light Microscopy and Image Processing Facilities > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > > http://www.bioimagexd.net BioImageXD > http://pacific.mpi-cbg.de Fiji - is just ImageJ (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Dresden Imaging Facility Network > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) |
 
|
Louis Villeneuve |
Re: Poor man's confocal-deconvolution
|
|
Hi to all, what would be the choice of algorythm for deconvolution for such a scope (apotome or optigrid)? I am a Huygens user and there is different adaptation (or algorhytm) according the type of scope your are using. Louis
Good summary. It depends on how brightly your samples are labeled and your resolution needs. But, for the additional US $20,000, you can purchase a lot of computing power to speed the deconvolution and gain a more flexible tool. That power can be applied to images collected by other modalities. It can also be put to work for analysis since deconvolution can produce quantitative images. You may want to investigate means by which decon may be speeded up, as in a different algorithm or implementation, improve the SNR of your images through optimizing the specimen (are specimen optics matched to the imaging system) and perhaps reducing the number of voxels that need to be deconvolved. eg. A full set of iterations may not be required for every channel. Regards, Glen Glen MacDonald Core for Communication Research Virginia Merrill Bloedel Hearing Research Center Box 357923 University of Washington Seattle, WA 98195-7923 USA (206) 616-4156 [hidden email] On May 27, 2010, at 2:03 AM, Daniel James White wrote: > Hi Bob, > > > On May 27, 2010, at 7:02 AM, CONFOCALMICROSCOPY automatic digest system wrote: > >> Date: Wed, 26 May 2010 11:32:12 -0700 >> From: Bob Nienhuis <[hidden email]> >> Subject: Poor man's confocal >> >> Anybody have any experience with the Zeiss Apotome structured illumination >> system? > > Yes, I works and does what its supposed to do pretty well. > However there are limitations, and one should not attempt to compare it closely with a point scanning confocal, > its not supposed to replace a confocal, rather its a very different approach, that happens to be relatively cheap. > > Apotome and Optigrid are basically rather similar.... > the limitations are: > 1) that you must take several (eg 3) images of each filed of view - so bleaching can become a problem > 2) if the lamp is unstable then some fo the 3 images are darker than others, this messed up the final image. > 3) One must carefully align the grid in the in its position conjugate to the image plane, > since you need to hgrid parrern to have as good contrast as you can get for the mthod ot work. > 4) very thick samples or samples with very few features will not work. > 5) the result can be a bit noisy (compared to the input images) , because the simple algorithm used throw away much of the signal. > 6) Dont try to measure relative fluorescence intensities ... the method is not linear. > good things are: > 1) Price, its simple and cheap > 2) Speed - pretty fast compared to point scanning a large field of vire > 3) no other special stuff needed - same stand and filters and objectives etc. > 4) Great for screening samples / yes or no answers. > > >> >> I am writing a proposal for a new microscope system and the vendor mentioned >> this as a way >> to reduce background in fluorescence microscopy. > > It works, > but there is, also quite cheap technology coming out now, which i think does a better job: > > Zeiss Viva Tome (3 colours) and the same gadget in a different box from Andor (4 colours possible apparently) > Method is called aperture correlation, and its like a mixture of spinning disk confocal and structured illumination > with collection of out of focus and mostly infocus signal, then an attempt to remove most of the residual blurr from the > moistly in focus signal my subtracting a weighted out of focus image. > > It works really well (I saw it live at ELMI 2010) and is faster then apotom/optigrid as only one CCD exposure is needed. > So less bleaching. > > I want one. I would say aperture correlation "seems better" than the older simpestructured illumination methods. > > Super resolution structured illumination methods like the Zeiss Elyra and Applies precision OMX > are a totally different story... dont confuse them with the apo tome/via tome stuff. > >> >> Apparently, it superimposes a moving grid on the image, and this somehow >> allows background >> noise reduction. > > look here for apotome > http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/apotomemechanics/index.html > and here for vivatome > http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/vivatomeopticaltrain/index.html > > aperture correlation method comes from Tony wilson's lab, see more references here > http://zeiss-campus.magnet.fsu.edu/referencelibrary/aperturecorrelation.html > >> >> We have used deconvolution to do this, but find that it takes a long time to >> do. > > its doenst have to take that long if you get a good feel for ot and have a computer with tens of processors. > > >> >> I think it would add about $20k to the cost of the scope. >> >> Opinions? > > apotome - good for what its designed for - but its not a poor mans confocal, its something rather different. > >> Worthwhile? > > definately > >> Any experience with it? > > yes, it works great for thin bright fixed samples. > > but vivatome looks better. > > > cheers > > Dan > > > >> >> Bob Nienhuis >> Neurobiology Research M/S151A3 >> UCLA / VA Medical Center >> North Hills, CA >> [hidden email] > > Dr. Daniel James White BSc. (Hons.) PhD > Senior Microscopist / Image Visualisation, Processing and Analysis > Light Microscopy and Image Processing Facilities > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > > http://www.bioimagexd.net BioImageXD > http://pacific.mpi-cbg.de Fiji - is just ImageJ (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Dresden Imaging Facility Network > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) |
|
|
Lutz Schaefer |
Re: Poor man's confocal-deconvolution
|
|
Louis,
the ApoTome is usually sold with the AxioVision
software, which (optionally for $$) also has a deconvolution module which can be
used to deconvolve the raw-datasets acquired with the ApoTome. The method used
for this modality is close to that of a regularized inverse filter. However, it
takes full advantage of an image formation model considering structured
illumination. As the OTF from such system does not significantly suffer
from a missing cone, the inverse filter gives very similar results to that of a
constrained iterative method, except the former is much faster.
Regards
Lutz
__________________________________
L u t z S c h a e f e r Sen. Scientist Mathematical modeling / Image processing Advanced Imaging Methodology Consultation 16-715 Doon Village Rd. Kitchener, ON, N2P 2A2, Canada Phone/Fax: +1 519 894 8870 Email: [hidden email] Website: http://home.golden.net/~lschafer/ __________________________________ From: [hidden email]
Sent: Thursday, May 27, 2010 1:03 PM
To: [hidden email]
Subject: Re: Poor man's confocal-deconvolution Hi to all, what would be the choice of algorythm for deconvolution for such a scope (apotome or optigrid)? I am a Huygens user and there is different adaptation (or algorhytm) according the type of scope your are using. Louis
Good summary. It depends on how brightly your samples are labeled and your resolution needs. But, for the additional US $20,000, you can purchase a lot of computing power to speed the deconvolution and gain a more flexible tool. That power can be applied to images collected by other modalities. It can also be put to work for analysis since deconvolution can produce quantitative images. You may want to investigate means by which decon may be speeded up, as in a different algorithm or implementation, improve the SNR of your images through optimizing the specimen (are specimen optics matched to the imaging system) and perhaps reducing the number of voxels that need to be deconvolved. eg. A full set of iterations may not be required for every channel. Regards, Glen Glen MacDonald Core for Communication Research Virginia Merrill Bloedel Hearing Research Center Box 357923 University of Washington Seattle, WA 98195-7923 USA (206) 616-4156 [hidden email] On May 27, 2010, at 2:03 AM, Daniel James White wrote: > Hi Bob, > > > On May 27, 2010, at 7:02 AM, CONFOCALMICROSCOPY automatic digest system wrote: > >> Date: Wed, 26 May 2010 11:32:12 -0700 >> From: Bob Nienhuis <[hidden email]> >> Subject: Poor man's confocal >> >> Anybody have any experience with the Zeiss Apotome structured illumination >> system? > > Yes, I works and does what its supposed to do pretty well. > However there are limitations, and one should not attempt to compare it closely with a point scanning confocal, > its not supposed to replace a confocal, rather its a very different approach, that happens to be relatively cheap. > > Apotome and Optigrid are basically rather similar.... > the limitations are: > 1) that you must take several (eg 3) images of each filed of view - so bleaching can become a problem > 2) if the lamp is unstable then some fo the 3 images are darker than others, this messed up the final image. > 3) One must carefully align the grid in the in its position conjugate to the image plane, > since you need to hgrid parrern to have as good contrast as you can get for the mthod ot work. > 4) very thick samples or samples with very few features will not work. > 5) the result can be a bit noisy (compared to the input images) , because the simple algorithm used throw away much of the signal. > 6) Dont try to measure relative fluorescence intensities ... the method is not linear. > good things are: > 1) Price, its simple and cheap > 2) Speed - pretty fast compared to point scanning a large field of vire > 3) no other special stuff needed - same stand and filters and objectives etc. > 4) Great for screening samples / yes or no answers. > > >> >> I am writing a proposal for a new microscope system and the vendor mentioned >> this as a way >> to reduce background in fluorescence microscopy. > > It works, > but there is, also quite cheap technology coming out now, which i think does a better job: > > Zeiss Viva Tome (3 colours) and the same gadget in a different box from Andor (4 colours possible apparently) > Method is called aperture correlation, and its like a mixture of spinning disk confocal and structured illumination > with collection of out of focus and mostly infocus signal, then an attempt to remove most of the residual blurr from the > moistly in focus signal my subtracting a weighted out of focus image. > > It works really well (I saw it live at ELMI 2010) and is faster then apotom/optigrid as only one CCD exposure is needed. > So less bleaching. > > I want one. I would say aperture correlation "seems better" than the older simpestructured illumination methods. > > Super resolution structured illumination methods like the Zeiss Elyra and Applies precision OMX > are a totally different story... dont confuse them with the apo tome/via tome stuff. > >> >> Apparently, it superimposes a moving grid on the image, and this somehow >> allows background >> noise reduction. > > look here for apotome > http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/apotomemechanics/index.html > and here for vivatome > http://zeiss-campus.magnet.fsu.edu/tutorials/opticalsectioning/vivatomeopticaltrain/index.html > > aperture correlation method comes from Tony wilson's lab, see more references here > http://zeiss-campus.magnet.fsu.edu/referencelibrary/aperturecorrelation.html > >> >> We have used deconvolution to do this, but find that it takes a long time to >> do. > > its doenst have to take that long if you get a good feel for ot and have a computer with tens of processors. > > >> >> I think it would add about $20k to the cost of the scope. >> >> Opinions? > > apotome - good for what its designed for - but its not a poor mans confocal, its something rather different. > >> Worthwhile? > > definately > >> Any experience with it? > > yes, it works great for thin bright fixed samples. > > but vivatome looks better. > > > cheers > > Dan > > > >> >> Bob Nienhuis >> Neurobiology Research M/S151A3 >> UCLA / VA Medical Center >> North Hills, CA >> [hidden email] > > Dr. Daniel James White BSc. (Hons.) PhD > Senior Microscopist / Image Visualisation, Processing and Analysis > Light Microscopy and Image Processing Facilities > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > > http://www.bioimagexd.net BioImageXD > http://pacific.mpi-cbg.de Fiji - is just ImageJ (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Dresden Imaging Facility Network > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) |
|
|
Daniel James White |
Re: Poor man's confocal-deconvolution
|
|
In reply to this post by Louis Villeneuve
Hi Louis,
Begin forwarded message: > > Date: Thu, 27 May 2010 13:03:21 -0400 > From: [hidden email] > Subject: Re: Poor man's confocal-deconvolution > > > Hi to all, > > what would be the choice of algorythm for deconvolution for such a scope=20 > (apotome or optigrid)? I am a Huygens user and there is different=20 > adaptation (or algorhytm) according the type of scope your are using. > it doesn't really make sense to deconvolve the result image from apotome or optigrid.... In a way its already deconvolved. Of course you could measure beads and use a PSF calculated from those, and you would probably get a small increase in contrast... but unless the original 3 images were super high signal: noise, the result aptoome/optigrid images arent as linear as you really required to do a deconvolution that makes sense. Lamp flicker could really mess the image up already, so they were never linear to begin with. I was recently informed of a paper showing optigrid result images can be linear in a low noise system, using a stable metal halide lamp imaging beads. see http://www3.interscience.wiley.com/journal/113474320/abstract?CRETRY=1&SRETRY=0 MICROSCOPY RESEARCH AND TECHNIQUE 70:76–84 (2007) I dont mean you can't do deconv on apotome/optigrid result images ... but you might not really know what the result image means. and it might not make mathematical sense to do that. I mean dont treat deconvolution as a magic black box.... you need to understand what you put in to understand what you get out. Dan Dr. Daniel James White BSc. (Hons.) PhD Senior Microscopist / Image Visualisation, Processing and Analysis Light Microscopy and Image Processing Facilities Max Planck Institute of Molecular Cell Biology and Genetics Pfotenhauerstrasse 108 01307 DRESDEN Germany +49 (0)15114966933 (German Mobile) +49 (0)351 210 2627 (Work phone at MPI-CBG) +49 (0)351 210 1078 (Fax MPI-CBG LMF) http://www.bioimagexd.net BioImageXD http://pacific.mpi-cbg.de Fiji - is just ImageJ (Batteries Included) http://www.chalkie.org.uk Dan's Homepages https://ifn.mpi-cbg.de Dresden Imaging Facility Network dan (at) chalkie.org.uk ( white (at) mpi-cbg.de ) |
|
|
Daniel James White |
Re: Poor man's confocal-deconvolution
|
|
In reply to this post by Louis Villeneuve
Hi Lutz and Louis,
Begin forwarded message: > > Date: Thu, 27 May 2010 19:59:07 -0400 > From: Lutz Schaefer <[hidden email]> > Subject: Re: Poor man's confocal-deconvolution > > Louis, > > the ApoTome is usually sold with the AxioVision software, which = > (optionally for $$) also has a deconvolution module which can be used to = > deconvolve the raw-datasets acquired with the ApoTome. The method used = > for this modality is close to that of a regularized inverse filter. = > However, it takes full advantage of an image formation model considering = > structured illumination. This functionality is not present in Huygens as far as i know.... maybe the SVI guys could implement this model of image formation to generate theoretical PSFs? At least the Zeiss rep i spoke to recently at ELMI said to a room of people that deconvolution from Apotome probably isnt sensible.... maybe i misunderstood. Maybe he was reffering to VivaTome. So I am interested to hear that AxioVision has a deconv function for apotome... I should look for that... > As the OTF from such system does not = > significantly suffer from a missing cone, the inverse filter gives very = > similar results to that of a constrained iterative method, except the = > former is much faster. ...but it the result linear enough to make quantitative intensity measurements within an image? What are the limitations here? cheers Dan Dr. Daniel James White BSc. (Hons.) PhD Senior Microscopist / Image Visualisation, Processing and Analysis Light Microscopy and Image Processing Facilities Max Planck Institute of Molecular Cell Biology and Genetics Pfotenhauerstrasse 108 01307 DRESDEN Germany +49 (0)15114966933 (German Mobile) +49 (0)351 210 2627 (Work phone at MPI-CBG) +49 (0)351 210 1078 (Fax MPI-CBG LMF) http://www.bioimagexd.net BioImageXD http://pacific.mpi-cbg.de Fiji - is just ImageJ (Batteries Included) http://www.chalkie.org.uk Dan's Homepages https://ifn.mpi-cbg.de Dresden Imaging Facility Network dan (at) chalkie.org.uk ( white (at) mpi-cbg.de ) |
|
|
Lutz Schaefer |
Re: Poor man's confocal-deconvolution
|
|
Dan,
due to the size, I sent you off-list some experimental result images. For the processing it is better to use the raw images as the processed sectioned image may have suffered too much loss in numerical detail due to re-discretisation after demodulation or reconstruction. The method used in the Zeiss software is patented and we are also in the process of publication. What the nonlinearity concerns: We quantified the differences between the nonlinear magnitude- and the linear demodulation of the images and found, that for low enough grating frequencies they become insignificant. The Apotome uses grating frequencies that typically lie around less than 10% of the lateral OTF cutoff, which fulfills the condition above. Regards Lutz __________________________________ L u t z S c h a e f e r Sen. Scientist Mathematical modeling / Image processing Advanced Imaging Methodology Consultation 16-715 Doon Village Rd. Kitchener, ON, N2P 2A2, Canada Phone/Fax: +1 519 894 8870 Email: [hidden email] Website: http://home.golden.net/~lschafer/ ___________________________________ -------------------------------------------------- From: "Daniel James White" <[hidden email]> Sent: Friday, May 28, 2010 4:50 AM To: <[hidden email]> Subject: Re: Poor man's confocal-deconvolution > Hi Lutz and Louis, > > Begin forwarded message: > >> >> Date: Thu, 27 May 2010 19:59:07 -0400 >> From: Lutz Schaefer <[hidden email]> >> Subject: Re: Poor man's confocal-deconvolution >> >> Louis, >> >> the ApoTome is usually sold with the AxioVision software, which = >> (optionally for $$) also has a deconvolution module which can be used to >> = >> deconvolve the raw-datasets acquired with the ApoTome. The method used = >> for this modality is close to that of a regularized inverse filter. = >> However, it takes full advantage of an image formation model considering >> = >> structured illumination. > > This functionality is not present in Huygens as far as i know.... > maybe the SVI guys could implement this model of image formation > to generate theoretical PSFs? > > At least the Zeiss rep i spoke to recently at ELMI said to a room of > people > that deconvolution from Apotome probably isnt sensible.... maybe i > misunderstood. > Maybe he was reffering to VivaTome. > So I am interested to hear that AxioVision has a deconv function for > apotome... I should look for that... > > >> As the OTF from such system does not = >> significantly suffer from a missing cone, the inverse filter gives very = >> similar results to that of a constrained iterative method, except the = >> former is much faster. > > ...but it the result linear enough to make quantitative intensity > measurements within an image? > What are the limitations here? > > cheers > > Dan > > > Dr. Daniel James White BSc. (Hons.) PhD > Senior Microscopist / Image Visualisation, Processing and Analysis > Light Microscopy and Image Processing Facilities > Max Planck Institute of Molecular Cell Biology and Genetics > Pfotenhauerstrasse 108 > 01307 DRESDEN > Germany > > +49 (0)15114966933 (German Mobile) > +49 (0)351 210 2627 (Work phone at MPI-CBG) > +49 (0)351 210 1078 (Fax MPI-CBG LMF) > > http://www.bioimagexd.net BioImageXD > http://pacific.mpi-cbg.de Fiji - is just ImageJ (Batteries Included) > http://www.chalkie.org.uk Dan's Homepages > https://ifn.mpi-cbg.de Dresden Imaging Facility Network > dan (at) chalkie.org.uk > ( white (at) mpi-cbg.de ) |
|
|
In reply to this post by Bob Nienhuis
Hi Bob,
The only advantage of the Apotome or OptigGrid is that it uses the same filter cubes that you look at the specimen by eye. I second the comments of many other replies - especially Glen MacDonald and Daniel James White - and add a few... 1. I find to get good images I need to average 5 acquisitions of the 3 grid positions - so 15 images are grabbed per filter cube per Z-position. Figure 0.3 seconds times 15 images (and ignoring overhead of averaging), ~4.5 seconds per channel per Z-position (ORCA-ER, unbinned). To minimize Axiovert 200M filter cube moves, data are acquired in the order of filter cube #1 Z-series, then cube #2, etc. 2. The Zeiss Axiovision software (with modules that can be purchased) has the ability to deconvolve a Apotome data set. It uses the raw data (the 15 images per channel per Z-position) so you do not see the standard Apotome result. On the ~2 year old PC (4 Gb ram), Zeiss deconvolution is slow using the deconvolution settings I originally used - like overnight. My Zeiss reps did show me alternative deconvolution settings, so it is more like an hour for a decent size data set. 3. I hosted a Zeiss On Your Campus (ZOYC) event in March and saw the Vivatome - I believe the same demo instruments were at UCLA two weeks later. Most of the demo users and I did the same thing: we spent most of our time with the Zeiss LSM780. Ask your UCLA colleagues for take on the VivaTome. 4. fyi - The Axiovision 3D rendering is very nice - a much less expensive and more useful purchase than an Apotome. On the other hand, ZEN LE has pretty good rendering and is free (thanks Zeiss!). Two suggestions: 1. consider getting a very fast PC, lots of RAM (Vista 64-bit), and find a great deconvolution program that your users find worthwhile using. Volocity and Imaris come to mind. 2. If you really want to buy an Apotome - how about buying the Apotome (hardware, software modules that were bought with it) that I manage? Ours is on an Axiovert 200M. If you want to try before you buy, come on over to smog free Miami for several days (and nights - South Beach is close by!). If someone else wants to buy our Apotome - come on down and try it out. The price I would be looking for is the good motorized stage (and Axiovision driver and stage tiling) that goes on the Axiovert 200M (about $10K). Sincerely, George At 02:32 PM 5/26/2010, you wrote: Anybody have any experience with the Zeiss Apotome structured illumination system? George McNamara, Ph.D. Image Core Manager Analytical Imaging Core Facility University of Miami, Miller School of Medicine Miami, FL 33136 [hidden email] [hidden email] 305-243-8436 office http://www.sylvester.org/AICF (Analytical Imaging Core Facility) http://www.sylvester.org/AICF/pubspectra.zip (the entire 2000+ spectra .xlsx file is in the zip file) http://home.earthlink.net/~geomcnamara |
|
|
Lutz Schaefer |
Re: [VirusGuard] Re: Poor man's confocal
|
|
In reply to this post by Bob Nienhuis
George,
I have a couple of comments to your recent
posting:
1) The term "averaging" you mentioned in the
software was unfortunately not a good naming choice by Zeiss. When this feature
is enabled, in your case there would be 5x3 = 15 grid
positions and not 3! The reconstruction which solves a linear
system of equations is then over-determined as you have more equations than
unknowns and the result being calculated is by least squares (without
additional overhead). If you still want to call it "averaging" its okay, but I
think you get slightly more information due to the many more
grid-positions.
2) The deconvolution in case of ApoTome raw images
is just a simple regularized inverse filter, that should be very fast compared
to constrained iterative methods. I assume, that your data sets are very large
for which case any deconvolution software would be slow. Perhaps there are
additional issues in the setup, that limited the resources on your
PC?
Sincerely
Lutz
__________________________________
L u t z S c h a e f e r Sen. Scientist Mathematical modeling / Image processing Advanced Imaging Methodology Consultation 16-715 Doon Village Rd. Kitchener, ON, N2P 2A2, Canada Phone/Fax: +1 519 894 8870 Email: [hidden email] Website: http://home.golden.net/~lschafer/ ___________________________________ From: [hidden email]
Sent: Friday, May 28, 2010 10:03 PM
To: [hidden email]
Subject: [VirusGuard] Re: Poor man's confocal Hi Bob, Anybody have any experience with the Zeiss Apotome structured illumination system? George McNamara, Ph.D. Image Core Manager Analytical Imaging Core Facility University of Miami, Miller School of Medicine Miami, FL 33136 [hidden email] [hidden email] 305-243-8436 office http://www.sylvester.org/AICF (Analytical Imaging Core Facility) http://www.sylvester.org/AICF/pubspectra.zip (the entire 2000+ spectra .xlsx file is in the zip file) http://home.earthlink.net/~geomcnamara |
|
|
Hi Lutz,
Thanks for letting me know that the Apotome feature is more sophisticated than simple averaging. How about re-writing the Zeiss documentation to make this clear. Yes, I acquire full frame unbinned (1344x1024 pixels) Z-series. The PC is running Windows Pro XP32, 4 Gb ram, Axiovision is usually the only application running. Sincerely, George At 10:45 AM 5/29/2010, Lutz Schaefer wrote: George, George McNamara, Ph.D. Image Core Manager Analytical Imaging Core Facility University of Miami, Miller School of Medicine Miami, FL 33136 [hidden email] [hidden email] 305-243-8436 office http://www.sylvester.org/AICF (Analytical Imaging Core Facility) http://www.sylvester.org/AICF/pubspectra.zip (the entire 2000+ spectra .xlsx file is in the zip file) http://home.earthlink.net/~geomcnamara |
«
Return to Confocal Microscopy List
|
1 view|%1 views
| Free forum by Nabble | Edit this page |