So far, the scientists, from biotech company Helicos Biosciences, have successfully sequenced the M13 virus, but they hope to demonstrate similarly promising results on the human genome in the future. Unlike other methods, the technique builds up the sequence of each strand of DNA one base at a time.

"It's the first ever report of being able to sequence single molecules of DNA," says Tim Harris, one of the researchers. "People have been trying to do this for a long time."

Now that scientists have achieved a reference genetic sequence of the human genome, they are looking for cheaper and faster ways of gaining genetic information about individual variations, to compare to the reference genome. Helicos Biosciences' technique, reported in Science, requires less manipulation and no amplification of the genetic sample, which is costly and can introduce errors into the sequencing data, according to Harris.

Harris's technique involves a number of steps to build up the genetic sequence. First, the DNA sample is segmented into single strands of roughly 100-200 base pairs in length. To the end of each fragment, the scientists add a poly(A) tail - a segment of 50 adenine bases. This sample is then washed over a glass cover slip coated in poly(T) molecules - a long segment of DNA consisting of thymine bases. The poly(A) tail hybridises with the immobilised poly(T) segments, binding the DNA segments to the glass and keep them in place.

The next few stages of the process build up the genetic sequence of the sample one base at a time. To do this, the different bases that make up DNA, each joined to a dye molecule, are each washed over the fixed segments in turn many times. The bases attach to the corresponding bases (from their base pair) on the DNA segments - building up a full double helix from the tail end of the segment with one base with each wash.

The dye indicates which segments bonded with the base, and an image is taken of the slide to record this. The dye is then washed from the segments and another, different base is added to the sample, and the process is repeated. After many different cycles, the scientists can determine the genetic sequence of each segment by recording the order in which the bases pairs built up on the single strand of DNA. Once each segment has been sequenced, the scientists can compare this to the reference sequence to decide where each part fits into the whole sequence.

Helicos Biosciences estimate that the cost of this process for a human genome is roughly $72,000 dollars, and it would take eight weeks to perform on just one machine. However, the Science paper demonstrated the technique on a virus genome, and the company still needs to tweak a few steps in the process before it can be applied to the human genome.