The nomadic pastoralists who overwintered at a semi-permanent settlement at Shanidar, in northern Iraq, some 9000 years ago, had similar tastes to modern consumers: they preferred tender young lamb to mutton.

The Shanidar site contained a high concentration of bones from one-year-old animals. Doubtless the sheep also contributed their warm fleeces to keep the farmers and their families warm against northern Iraq's bitterly cold winters.

Shanidar has provided some of the earliest evidence for the domestication of sheep, the first wild animals domesticated as livestock. Mitochondrial DNA analyses points to at least three independent domestication events across Eurasia that began as early as 11,000 years ago.

The people of Shanidar, and other unknown shepherds in Europe, the Middle East, and possibly in Asia, began a selection process that is recorded in the DNA of today's sheep breeds.

The international Sheep Genome Project will throw light on the origins and evolutionary relationships of modern breeds, and explore the full extent of the genetic diversity that ancient pastoralists captured during the domestication process.

The International Sheep Genomics Consortium (ISGC) completed a first pass sequence of the Ovis aries genome in January. The consortium, which includes researchers from across the world, has been taking the lead in developing genomic tools and reagents for use by the sheep genomics community.

Dr James Kijas, leader of CSIRO Livestock Industries sheep genome group, says the group was interested to see if the sheep genome could be sequenced using a new short-read technology like Roche's 454 FLX sequencers.

"The emergence of much higher throughput and cheaper sequencing methods has helped us tremendously," Kijas says.

The consortium decided to invest in using Roche 454's FLX automated sequencers in two locations, New Zealand's University of Otago and the Human Genome Sequencing Center at the Baylor College of Medicine in Houston Texas.

The 454 methodology generates reads which are shorter in length than traditional Sanger reads. However, the ability to generate 60 million base pairs of sequence and 250,000 reads in a single instrument run has opened the door to producing genome sequences much more quickly and cheaply.

Kijas says the approach was driven by the need to characterise the variable component of the sheep genome, so the project had sequenced DNA from six different sheep breeds.

Researchers in Texas sequenced the DNA of three of the animals, while the team in New Zealand sequenced the other three.

Dr John McEwan, a senior scientist from AgResearch who led the New Zealand component of the work, says the approach was different to that employed in other genome projects. "From the outset, we used next-generation sequencing technology, and we've done things in a slightly different order.

"The previous strategy was to do a reference sequence, and then use that as the basis for locating SNPs (single-nucleotide polymorphisms). In our case, we used the bovine genome as a template, because it's the nearest available neighbour to sheep among livestock breeds sequenced.

"We sequenced six individuals to identify SNPs, then roughly positioned them using the bovine genome as a reference.

"We intend in the longer term to proceed towards a sheep reference genome, but the primary driver is the need to identify SNPs for a SNP chip."