Do-it-yourself gel imaging system
by Gheorge Christol, Labtimes 02/2011
In the autumn of 2009, the Gel documentation system used in Gheorge Christol’s lab completely broke down and he was assigned to find a cheap replacement. He had two options: either buy a new commercial system for a few thousand euros or acquire a used one online and hope for the best. He ended up going with a third solution – building a home-made gel documentation system.
A commercial Gel Doc system consists of a camera and lens for image acquisition, a computer and display for control and processing, and a software suite for quantifying or adjusting the images. Sometimes, Gel Doc systems have a built-in thermal printer (which is just a sneaky way of making you buy their expensive paper). The weakest point of any commercial Gel Doc system is the camera; usually nothing but a cheap point-and-shoot camera, with a mediocre lens and small pixel size (the ones with good cameras cost a lot). Their strongest point is the integration of image acquisition with image processing and band quantification. Below, I describe the basic design of my homebuilt Gel Doc system and some of the compromises and thoughts that went into it.
Figure 1 (left): Enclosure Diagram, Top View. A,B: holes drilled for screws that hold the L-bracket. C: footprint of the L-bracket. D: the cut-out for the lens (73mm diameter)
Figure 2: Schematic view of camera mounting using an L-bracket. A: Top portion of the GelDoc enclosure. B: L-bracket (10-15cm long). C: dSLR Camera. D: Screw that goes through a centre hole in the L-bracket and fits into the camera tripod socket. E: Screw and nut (two of each) that are used to attach the L-bracket to the enclosure.
(Figures: Gheorghe Christols)
Homebuilt Gel Doc system Components:
- Camera and Lens: any Canon dSLR that has live-view and its kit lens.
- Camera enclosure and platform.
- Computer: any Mac or Windows XP, Windows Vista or Windows 7 computer with a USB port.
- Software: free Canon dSLR software (acquisition) and ImageJ (open-source image processing programme).
All modern dSLR cameras have sensors that are several times larger than point-and-shoot cameras and are therefore well suited for a homebuilt Gel Doc. As of winter 2011, only Canon ships their dSLRs with free remote capture software. This is software that allows the user to control the camera from a computer via a USB cable.
It is important that the camera has “Live-View” capabilities (Canon marketing lingo) because this allows the user to adjust focus/exposure from the computer while viewing a real-time video stream from the camera. Some of the older generation cameras don’t have the Live-View feature and are therefore much less useful.
These requirements are satisfied by several Canon cameras like EOS 400D, 450D, 500D, 550D, etc. (In the US these cameras are known as Rebel XS, Rebel XSi, Rebel T1i, Rebel T2i, Rebel T3i, Rebel T3, etc). The cheapest of the bunch can be bought online with the kit lens for as little as €500. If you feel like hunting for bargains, you can try buying a refurbished or even used older model (such as EOS 1000D [Rebel XS] and EOS 450D [XSi]) for €200 to €300.
The 18-55mm kit lens will be fine for all but the most demanding users. If you want a faster lens you can buy a 50mm f/1.8 lens for around e100, but you will lose the zoom capability and you will have to construct an enclosure tall enough to give you a wide enough field of view with the 50mm lens.
Camera mounted on the darkroom cabinet and connected with data and power cable. (Photo: Gheorghe Christols)
You will need a colour filter to prevent UV light from reaching the camera sensor. Here, I am assuming that most people use the Gel Doc system to image their Ethidium Bromide or SYBR Safe gels (the most common DNA dyes, they absorb in the deep blue/UV range and emit in the orange/red range). The most convenient, commercially available filters are Tiffen Orange 16 or Tiffen Orange 21 glass filters. The 58mm diameter filters cost around €15 new (58mm fits the 18-55mm kit lens that comes with most Canon EOS cameras).
In addition, you will need an AC/DC power supply, so you can keep the camera plugged in all the time and not worry about charging batteries. You will also need a USB cable long enough to go from the computer to the camera, with a bit extra left over.
This one can be as simple as a thick cardboard box on which you place the camera. If you want longevity, I would recommend making a metal, wood or plexiglass enclosure. Our enclosure was made of 5mm-thick black plexiglass by the chemistry department machine shop. It has a door and a cover-top but those are not essential. We designed it so the enclosure fits on the SYBR Safe transilluminator, UV transilluminator, as well as the white light table.
If you want to make sure nobody walks away with your camera, you can buy an adhesive plate that you can glue to the side of the camera and a steel cable so you can secure the camera to the work-bench. Our enclosure is cubic with 30 centimetres on each side. It has four walls and a top surface with a 73mm-diameter cut-out in the centre to accommodate the lens.
After opening the door of the gel documentation system, gels may be laid on the transillumination table and captured by the camera. (Photo: Gheorghe Christols)
You can use either Mac or PC. We are using an old Windows XP machine that is quite slow but is still enough for this purpose. If you’re buying a new computer, you can opt for a low cost desktop or a so-called NetTop PC. Be aware that the keyboard and mouse will have traces of Ethidium Bromide and you need to wear gloves when using this PC.
The bare minimum is Canon EOS Utilities (comes with the camera) and ImageJ (open-source). EOS Utilities has a remote-capture component, where you can view the real-time feed from the camera sensor and control camera settings. I recommend setting the camera in the manual mode, BW image mode, no sharpening and no contrast adjustments. Set the lens focus to manual on the lens barrel. You will have to adjust focus from the software while monitoring the live-view. You may have to adjust the focus while using visible illumination because the UV image may be too faint to view clearly in live-view. You might need to set the exposure time to 1-2 seconds, aperture f/5.6-f/8 and ISO 100-400. If the gel requires longer exposure, feel free to crank up the ISO to 1600 or even 3200 if needed. You might get slightly noisier images but they’re still quite good. Don’t overexpose because you will be clipping highlights and losing information about the brightest bands.
Once you have acquired the image from the camera (JPEG is recommended for simplicity), you can open it in ImageJ, and adjust the rotation angle and crop area. You can also invert the image and make basic brightness/contrast adjustments. ImageJ has an option where you can quantify the DNA/Protein content in gels (as long as you have control lanes). Note that the homebuilt Gel Doc system will not have a strictly linear response (i.e. you can tell, whether a band has more DNA than another and you can even roughly estimate how many times more DNA it contains but its response curve is not linear). Make sure that you use linear image adjustments in ImageJ (i.e. set the white point, set the black point but don’t use s-curves for image adjustment because s-curves distort relative band intensities).
You can build this Gel Doc system together for under €1,000 in less than a week. It is recommended that you have the camera enclosure/platform made by someone from a machineshop. A well-made enclosure will outlast the camera. These cameras have a life expectancy of over 100,000 shutter actuations. Commercial Gel Doc systems are probably a bit more convenient because of the integration feature but this homebuilt contraption yields very good gel images, too, for less money. You can read detailed instructions for setting up the remote capture software at http://panic.berkeley.edu/~ghe/GelDoc/UsingGelDoc/ and get guidance for using ImageJ with the Gel Doc at: http://panic.berkeley.edu/~ghe/GelDoc/UsingImageJ/
(Gheorge Christol, UC Berkeley, Molecular Biophysics, Bustamante Lab)
Last Changed: 10.11.2012