Posted by
George McNamara on
URL: http://confocal-microscopy-list.275.s1.nabble.com/General-question-Software-vs-hardware-tp7589316p7589321.html
*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopyPost images on
http://www.imgur.com and include the link in your posting.
*****
Hi Mika,
White et al 1987 (
http://jcb.rupress.org/content/105/1/41.long ) made a
compelling case for point scanning confocal microscopes: collect just
the in focus light with instant gratification. The case has not changed,
the hardware (especially data deluge side) has gotten a lot better. I
note that both widefield detectors and PMT/APD/hybrid detectors/others
have gotten a lot better in the 32 years from 1987! As have the optics
and automation.
Paul Goodwin 2014 (
https://www.ncbi.nlm.nih.gov/pubmed/24974028 ) made
a nice case for quantitative deconvolution microscopy (and on very high
quality specimens, ~10% improvement in resolution compared to simple
widefield), but historically slow.
Now, with 'instant gratification' spatial deconvolution, thanks to the
GPU revolution (NVidia RTX Titan~16 Teraflops [S.P.], 24 Gb ram, $2500
... not including the deconvolution software module price), widefield,
spinning disk (and slightly exotic variants like iSIM, DMD based, etc,
see also new THUNDER Imagers [see p.p.s.]), multiphoton (I'm excited
about the price point of recent fiber lasers, which could become much
better price if achieve 'economy of scale'), and of course, point
scanning confocal microscopes.
Spatial deconvolution (especially if someone $uccessfully implements
joint spatial deconvolution and spectral unmixing, multiple cameras -
for 4 cameras see Babcock 2018, mentions aiming for 8 cameras) helps
with Expansion Microscopy and/or DNA-PAINT, to go super-resolution ...
really single molecule counting (and DNA-PAINT eliminates the classic
issue of PALM/STORM/FPALM of not counting every molecule). Sure,
DNA-PAINT (like STORM etc) have the issue of a whole lotta images
acquired. Data deluge: who cares? Jerome & Price's 10th commandment of
confocal imaging is: "10. Storage Media Is Essentially Free and Infinite".
More significantly, DNA-PAINT and related methods (single molecule RNA
FISH, scRNAseq -> MERFISH = Moffitt 2018 as example, etc) also enable
multiplex -- with single molecule counting -- to whatever plex is needed
to answer the 'biological question(s)' being posed.
All that said, the installed base of research grade point scanning
confocal microscopes is large (5000+) and efficient at acquiring high
quality images, to the point that user's sample preparation (and
avoidance of purchasing stuff from 'Santa Crap' and similar companies)
is much more limiting than the microscopes.
George
p.s. a couple of references not included in above:
W. Gray (Jay) Jerome, Robert L. Price 2018... Basic Confocal Microscopy
second edition
https://link.springer.com/book/10.1007%2F978-3-319-97454-5Expansion ... X10 protocol ... Truckenbrodt 2019 Nat Protoc,
https://www.nature.com/articles/s41596-018-0117-3DNA-PAINT acronym soup review ... Nieves 2018 Genes,
https://www.mdpi.com/2073-4425/9/12/621Babcock 2018 (4 --> 8 cameras, single molecule localization microscopy
with $1550 CMOS cameras) ...
https://www.nature.com/articles/s41598-018-19981-zMoffitt et al 2018 ...
http://science.sciencemag.org/content/362/6416/eaau5324.long and
commentary
http://science.sciencemag.org/content/362/6416/749***
Some resolution numbers:
1.4 NA objective lens ... 500 nm wavelength .. dxy = 0.61 * wqavelength
/ NA (I routinely drop the 0.61 from 0.61)
0.6 * 500 / 1.4 = 214 nm
widefiel deconvolution (re: Goodwin 2014) ~10% better ... 193 nm (if
pixel size matched or interpolate optimally).
point scanning confocal -- Zeiss has a nice PDF, "Zeiss 2008 Principles
- Confocal Laser Scanning Microscopy" (see fig 10) on confocal
resolution wrt 1 and smaller pinhole size, source of the values below,
1 Airy unit: 0.51 * 500 / 1.4 = 182 nm.
0.5 Airy unit ... 0.44 * 500 / 1.4 = 157 nm ... ~0.25 photons throughput
(which doesn't matter if target is photostable).
0.2 Airy unit ... you can ask your Zeiss rep about AiryScan (and
FastAiryScan).
0.1 Airy unit ... 0.37 * 500 / 1.4 = 132 nm ... ~0.10 photons throughput.
Most modern point scanning confocal microscopes have a 405 nm laser, so
if using BV421 (and ignoring potential photobleaching for a moment),
1 Airy unit: 0.51 * 421 / 1.4 = 153 nm.
or in reflection mode, i.e. nanodiamond or nanogold,
1 Airy unit: 0.51 * 405 / 1.4 = 147 nm ... and reflection implies no
photobleaching, so infinite number of photons (though also no blinking,
so not usually eligible for precision localization) ...
0.1 Airy unit: 0.37 * 405 / 1.4 = 107 nm
and not going completely exotic with NA (i.e. 1.65), if perfectly
refractive index match with a fairly conventional 1.49 NA lens, and
inreflectance:
0.1 Airy unit: 0.37 * 405 / 1.49 = 100.57 nm
I think I'd rather invest a DNA-PAINT friendly rig than deal with 157 to
101 nm.
DNA-PAINT makes resolution irrelevant, if you use it (and don't run out
of disk space or time or money), since precision localization is
resolution divided by square root of number of photons, ex. 250 nm XY
resolution / sqrt(1,000,000) = 0.25 nm, and could increase number of
photons per target further, but why bother?
***
point scanning confocal microscopes are also great platforms for
F-Techniques, such as FastFLIM (aka rapidFLIM, etc, much faster than
classic TCSPC slow FLIM), FCS, FCCS, see Liu 2008,
https://www.ncbi.nlm.nih.gov/pubmed/18387308***
p.p.s. Disclosures I am ...
1. currently hosting a Leica THUNDER Imager tour event (ends Monday
3/18/2019 afternoon) ... see pdf download page,
THUNDER Technology Note
THUNDER Imagers: How Do They Really Work?
https://www.leica-microsystems.com/science-lab/thunder-technology-note2. hosting Nikon confocal demos in May 2019.
3. aiming to co-host with ISS a FastFLIM (one day) mini-symposium this
summer.
4. an unpaid advisor for Gary Brooker for FINCH/CINCH, re:
https://www.ncbi.nlm.nih.gov/pubmed/28261321On 3/17/2019 10:43 AM, Mika Ruonala wrote:
> *****
> To join, leave or search the confocal microscopy listserv, go to:
>
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy> Post images on
http://www.imgur.com and include the link in your posting.
> *****
>
> Hi.
> There are software solutions that are able to create image data from wide-field and even microscope systems with seemingly similar quality to that obtained from confocal systems.
>
> The comparison of acquisition vs. software is essentially a comparison of image acquisition vs. image processing. While a software solution is way cheaper than a hardware solution if it is able to produce image data with equal quality why would anyone choose to invest to a confocal anymore?
>
> I’m looking forward to a vidid discussion!
>
>> m