SCIENCE AND VISUALIZATION

Mapping the Genome with a Microscope:
Optical Mapping at the Keck Lab

Edward J. Huff and Estarose Wolfson

[Ed: Links to web pages and/or e-mail addresses which have become inactive since the publication of this article have been enclosed in curly brackets { }. Replacement links have been provided where possible.]

The larger red grid (click on it, as on the other images here, for a larger view) is a collage of 8100 microscope images showing thousands of DNA samples that have been arrayed on a single surface. The blue images detail a small portion of a single such microscope image. The side of the red grid represents about 8 mm, and the bottom edge of each blue image represents about 30 microns (0.03 mm).

All of the images are captured on a sophisticated type of optical microscope known as a fluorescence microscope. Typical optical microscopes form images by shining light through a sample and magnifying the shadow of the object. Fibers as thin as one micron can easily be seen. However, DNA is only about 0.002 micron thick, much too thin to cast a shadow. The trick is to arrange for the DNA to emit light so that it stands out. This is done by staining it with a fluorescent dye. In the resulting digital images, 15 pixels represent about one micron, and the edge of one full-size image represents about 100 microns.

The purpose of making images of DNA is to find the location of specific landmarks along the DNA. One important class of enzymes in the molecular biologist's toolkit comprises the restriction enzymes. These enzymes cut the DNA wherever they find specific short sequences, and there are thousands of different enzymes available. If the DNA is under tension when the enzyme cuts, the DNA pulls away from the cut sites. This produces a visible gap in the image of the DNA molecules, which serves as a landmark that can be mapped. Measuring the location of these landmarks with a microscope is called optical mapping.

In our procedure, the glass surface that holds the DNA for optical mapping is chemically modified. When a small droplet containing thousands of identical DNA molecules is placed on this surface and allowed to dry, the molecules are elongated by fluid flow and stick to the surface, as shown in the blue images. A machine places many of these droplets at fixed grid positions on the glass surface -- in our example, 100 droplets on a 10x10 grid. Each droplet may contain DNA from a different source.

In order to expose landmarks along the DNA molecules, the surface is treated with a restriction enzyme. The enzyme cuts each DNA molecule at each restriction site, and the stretched DNA springs back. When this surface is treated with the fluorescent dye, mounted on the microscope, and scanned by an electronic camera, the resulting images look like those shown here. These images are pseudocolored for publication.

The green grid shows the actual spacing of the 81 individual images captured from one of the droplets; in the red grid, the spaces are eliminated. The bright specks in the two grids are the DNA molecules. Software developed in the lab automatically identifies each molecule in the image by isolating it from the background, and finds its central path or backbone -- shown here as a line of red pixels. Then the intensity levels along the backbone are analyzed, forming a profile of the molecule and identifying the locations of the cuts that serve as landmarks on a map of the DNA (shown in the graph at right).

The whole process is repeated using different restriction enzymes. Since each enzyme cuts at a different DNA sequence, using another enzyme exposes a whole new set of landmarks. In time, when enough cuts have been correlated, the entire sequence can be deduced.

Optical mapping is a new approach to the problem of mapping genomes. It directly produces ordered restriction maps of DNA samples, using small amounts of material, in a process that is being automated. Population genetics requires the analysis of the differences among many individuals. Optical mapping will provide a rapid and low-cost way to determine these differences.

Posted 25 September 1996. Last revised 20 May 2004.