Fluorescent microscopy pictures - need advice

Starting out in microscopy? Post images and ask questions relating to the microscope and get answers from our more advanced users on the subject.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

Starshade
Posts: 133
Joined: Wed May 04, 2011 1:47 am

Fluorescent microscopy pictures - need advice

Post by Starshade »

Hello everyone,

Recently I've been trying to take pictures in an axiostar at my lab, well I've been taking pictures for years and mostly it was obvious for me what to do, but when it came to fluorescence I got stuck.

Basically, I have axiostar microscope, x63 fluorescent and x100 oil imm lenses that I use and various fluorescent dyes like propidium iodide, syto and acridine orange. The camera is Canon PowerShot G10, it's not an SLR but it makes RAWs and has dSLR like interface with easily adjustable iso (80-3200) and shutter speed (up to 15 seconds).

I've been trying a lot of options but stopped with 80 and 1/4 - 1/2 seconds, at first I tried higher sensitivity

So far the best I've got
Image
This picture I just made for fun

Image
Acridine orange colored a ciliate

Image
The ciliate cysts appear like I used lots of photoshop effects, perhaps the effect is because I cropped small part of the original image

Image
Combination of two pictures but these two bunches of cysts were really close to each other on the sample

The cysts are roughly 5-8 microns in diameter, sorry for not adding bars

In terms of photoshop I edited the initial RAW files adding blacks, enhancing clarity and selectively increasing brightness and then applying little curves modification.

So the questions I have are:


Are there any secrets of taking good pictures of fluorescence aside from having better cameras ?

Is the photoshop editing that I did considered to be "cheating"? May be for publishing in a journal it's unacceptable but fine for conferences

How not to have less noise? Higher isos make the pictures really crappy and the details inside the cells are invisible

Did you ever have problem that pictures appear differently in various browsers? For example what I see on my screen in firefox and in IE or chrome are completely different. Basically, firefox shows what I see in my editors and IE and Chrome enhance noise and white noisy background appears around bright objects on the dark background (basically the same effect if you enhance exposition)

Thank you

Pau
Site Admin
Posts: 6053
Joined: Wed Jan 20, 2010 8:57 am
Location: Valencia, Spain

Post by Pau »

Sorry, I haven't experience in fluorescence, but about photogragraphic questions:
- Both high ISO and long exposure (and underexposed image if present)increase the noise. This effect is much higher in small sensor cameras. Noise suppression software equalizes noise at the cost of blurring detail. You are right shotting at base ISO
- Some internet browsers do not manage color profiles and show the images litke if them were taken in sRGB. The best approach will be to shot in sRGB o to convert the image in PS to sRGB before uploading it.
Pau

Starshade
Posts: 133
Joined: Wed May 04, 2011 1:47 am

Post by Starshade »

Thanks you, I learnt about color profiles and managed to convert my images into sRGB and now they are displayed the same way in different browsers.

At least one issue is resolved (didn't reupload pictures here)

Next time taking pictures I will try to decrease shutter speed and make it as low as I can and enhance light in PS at RAW stage and see what difference it makes.

Cactusdave
Posts: 1631
Joined: Tue Jun 09, 2009 12:40 pm
Location: Bromley, Kent, UK

Post by Cactusdave »

A major problem with fluorescence microscopy is the 'diffuse out of focus glow' problem caused by the contribution of fluorescence of structures above and below the plane of focus. This contributes to the difficulty in getting clear photographs of fluorescent subjects. One is always balancing between exposures which adequately define the fluorescent subject we are most interested in, and exposures which minimise the contribution of the out of focus glow. Some compensation for this can of course also be made in post processing using 'Levels' for example. The flatter the cell the less out of focus glow will be a problem, but for highly three dimensional cells it is a very serious issue.

It was in response to the problem of out of focus glow limiting the value of the increasingly important technique of immunofluorescence that the technique of confocal fluorescence microscopy was developed. If you are technically minded you may find this paper by Brad Amos and coleagues describing the first results with this now routine and immensely powerful technique of interest. http://www.ncbi.nlm.nih.gov/pmc/article ... 105141.pdf
Leitz Ortholux 1, Zeiss standard, Nikon Diaphot inverted, Canon photographic gear

Starshade
Posts: 133
Joined: Wed May 04, 2011 1:47 am

Post by Starshade »

Thank you for the info

I knew about the confocal method and "almost" used it once in my research. Was good to learn about out of focus glow.

The only problem with confocal microscopy, the images produced are very valuable in scientific terms but they lack quality of dSLR camera produced images.

By now I figured how to manage relatively large objects, it was about correct exposure indeed. I also noticed my 40x lens gives way better quality than 100x. At 40x I even managed to stack up to 10 images and soon I'll post some really nice results.

Cactusdave
Posts: 1631
Joined: Tue Jun 09, 2009 12:40 pm
Location: Bromley, Kent, UK

Post by Cactusdave »

I think that is a rather common observation with fluorescence microscopy. When I was doing this professionally the objectives I used most were x40 and X63. X100 never seemed to give particularly worthwhile additional resolution. High numerical apertures really help with the rather weak illumination produced by fluorescence so planapochromats were the order of the day, not necessarily good news for the amateur spending his own money. :cry:
Leitz Ortholux 1, Zeiss standard, Nikon Diaphot inverted, Canon photographic gear

discomorphella
Posts: 607
Joined: Sun Oct 01, 2006 7:26 pm
Location: NW USA

Post by discomorphella »

Assuming you're using the standard epi-illumination, the brightness of your fluorescence image will be roughly proportional to (NA)^4/(mag)^2. So a 40X/1.3 NA lens will be much brighter than a 100X/1.3 lens. This is why you usually see 60X/1.4 objectives used for epifluorescence images. I use my 40/1.3 and 63/1.4 for that exact reason, seldom resorting to the 100X. This is also assuming you have a sufficient number of pixels to properly record your image with the lower mag/higher NA objectives. Your SNR may improve just by a simple lens swap.
If you don't have a confocal setup, one way to remove out of plane fluorescence is to use 3D deconvolution. There are several good commercial programs (Autoquant, now Media Cybernetics, and Huygens are both good but expensive) out there which can greatly help, and there are also several imageJ plugins (free) which implement most of the available algorithms. I've used some of the 2D and 3D maximum likelihood algorithms and gotten very good results. You need to take a stack of images and you'll need a good workstation to get results quickly, but almost everyone uses some form of deconvolution for their widefield fluorescence images. Then of course there's 2-photon, 4-and 5-pi imaging and STED, PALM etc...more exotic techniques that require a more elaborate setup than your axiostar. There are some good review articles which I'll edit in references for.
As for what's acceptable in publication, most journals (Nature for example) have whole sets of guidelines on what constitutes acceptable image manipulation. Best consult the editor for where you're planning on submitting.

David

Starshade
Posts: 133
Joined: Wed May 04, 2011 1:47 am

Post by Starshade »

Thank you guys for your replies.

I think I will try ImageJ plugins to stack images, so far what I get is enough for current scientific purposes.

Using my limited resources I'm trying to just get better at microphotography and do it mostly for fun at this point but I know I will need this knowledge in the near future.

The other reason I use 40x more than 63x or 100x is because I want to get both phase contrast/DIC and epifluorescence images and then see how they match and my 63x lens is not PhC, and 100x... well the fluorescent picture is not good

Post Reply Previous topicNext topic