I am a scientist working in the genetics department in my institute. I have been using Gel Documentation Systems for some time now. The basic area of my work involves DNA gels & protein gels.

Most of the imaging that I do is under transilluminated Ultra Violet (UV) light & transilluminated white light.

All along, we have been using systems that provide a monochrome CCD camera (possibly analog) connected to a PC & we integrate our exposure times through the computer. This camera can be directly connected to an analog thermal printer (only) too & the output obtained.

Lately however, I have been hearing of Digital Cameras & some companies like Kodak have been marketing it saying that I can get three times the resolution. However, the established companies in this field (like BioRad, USA) say that the analog monochrome camera is best suited to my applications because it reads the output in 256 gray scale levels and provides more clear differentiation between the bands that I image. They also claim that the output that I get with color digital cameras like the Kodak DC 290, is actually one third of what is claimed because since the pixels are divided between RG&b (red, green, blue), every pixel has an assigned value & what I get os interpolated resolution. The cost of these systems is more than the digital camera based systems.

What do you suggest is the system that I should buy. Could you make the difference clearer to me between these 2 camera. As of now, I am pretty confused.

-- Vikram Singh (syscons@bol.net.in), March 19, 2000

Answers

Kodak technical support should be able to give you some idea of whether their system would work for you. What is the resolution you are getting now Depending on the size of the gels you are getting an alternative would be to look at using some of the desktop scanner technology to provide results for you. For example, if you placed a gel tube on a desktop scanner with a transparency adaptor, you could get a high resolution image of the gel patterns. You could probably substitute a UV source for the light in the transparency adaptor and get a UV image as well. Check this source for digital uv photography: http://www.foto.no/nikon/index2_PC.html. Hope this helps

Jonathan

-- jonathan ratzlaff (jonathanr@clrtech.bc.ca), March 19, 2000.

For what it's worth. The information you've been given about digital colour cameras is slightly misleading. It's true that each pixel only gives one channel of colour information, Red, Green, or Blue, but every pixel is used to give brightness information. If you visualise that the image is made up of two layers, a high-resolution brightness layer, overlayed with a lower resolution colour layer, that's more or less how it works. So the camera can see true detail down to the pixel resolution level, but the colour information is a "fudge", made up from an area of the surrounding pixels. I would say that the digital colour camera may well give you better resolution, as long as colour information is not important.

To make matters even more confusing, the pixels aren't even equally divided between the Red, Green and Blue channels. Usually more pixels are devoted to the Green channel than to the other two.

One potential problem that hasn't been mentioned is noise in the darkest levels of the image. Scanning Monochrome cameras can be made to have a very good performance in this area, but consumer colour cameras are made down to a price, not up to a specification, and can have a questionable noise performance.

I can only suggest that you try to persuade Kodak or another company to let you have one of their cameras on approval, and assess its performance alongside your existing system.

The best of luck.

-- Pete Andrews (p.l.andrews@bham.ac.uk), March 20, 2000.

If you're making quantitative densitometry measurements from your gel images, then a consumer digital camera may not be the way to go. (Or even a professional digicam, for that matter, without special software.) The problem is that they have a logarithmic relationship between image brightness and brightness values in the file, and their exposure systems mess about a fair bit with the tone curves. Thus, if you're trying to quantitate the amount of a protein, DNA fragment, etc by measuring area under a density curve, you'd get inaccurate results with consumer digicams. OTOH, if you're just measuring relative positions of things (as in sequencing), you'd probably do fine. A caveat with even the latter though: Most digicam lenses have a fair bit of geometric distortion that could affect the positional accuracy. If you had a standard setup (same focal length, subject-lens distance, etc) every time, you could probably come up with a standard correction using something like Photogenetics, using their "dewarp" plug-in. (See our review at http://www.imaging-resource.com/SOFT/PGEN/PGEN.HTM.) Good luck!

-- Dave Etchells (detchells@imaging-resource.com), March 20, 2000.

You have not mentioned what resolution you desire. If all you need is 100x100 pixel resolution, than a 2000x1500 digital camera's output will surely be adequate, no matter what r/g/b splitting/interpolation is involved. If you need precisely 2000x1500 resolution, then the hair-splitting becomes relevant. So what resolution do you really need? At a glance documenting gel-migration spots might not need super resolution, unless you are trying to image real small spots from a distance, or many at once.

When you test a quality camera like the Nikon Coolpix, etc be sure to put the thing into MONOCHROME mode, if all you want is luminance (brightness) data. Then you will have a better chance of getting resolution without color-interpolation artifacts.

Also be thinking of convenience factors. If all you want is television-resolution output (512x350) into a VCR or thermal printer, you don't even need to click the shutter on many digital cameras. You would just plug the analog output port of many digicams into your VCR or record away. If you want true digital picture frames, pick a unit that will be easiest for you to automatically control remotely with a computer. Some of the Kodak cameras come to mind as having all kinds of ways to control them via software on a PC.

And you're missing out on an obvious new development, brought to my attention courtesy Imaging-Resource. For max quality, consider waiting a few weeks for the fascinating new JVC digicam to hit the market. It has a feature where it will take three slightly offset exposures, and then combine them into one full-color less-interpolated result. It also has a neat mode where it takes three different exposures and combines them into one final image with a large, linear range of values from light to dark. If I had to do precise work I'd sure be contacting JVC and asking them when they're going to start shipping.

-- Russell Bozian (finaldesign@hotmail.com), March 20, 2000.