It is the sequel to one of the most astonishing scientific achievements of the past century and a technical tour de force involving the cataloguing of 4.9 trillion letters of human DNA code – enough to fill 300,000 copies of War and Peace.

Ten years on from the White House press conference at which President Clinton hailed the completion of the "first draft of the human book of life", scientists are now well on their way to compiling a substantial library.

Researchers have unveiled results from the pilot phase of the 1000 Genomes Project, which aims to sequence the genetic code of 2,500 people. It is the most complete inventory yet of the millions of variations between people's DNA. The details will illuminate the genetic roots of diseases, focus the search for treatments, and shed light on human evolution.

The $120m (£75m) project, involving hundreds of scientists in an international collaboration of universities, charities and companies, is using advanced gene-sequencing technology to map out the full diversity of human DNA. It has already examined the DNA of more than 800 people thanks to improvements in sequencing technology since the first complete draft of the human genome was published in 2000.

The research is published in a set of papers in Nature and Science.

After the first draft human genome sequence was published in 2000, the main focus of human genetics research has been to identify which differences between people's genome sequence, or genetic variants, contribute to disease. "In the last 10 years, DNA sequencing technology has advanced dramatically so that it becomes feasible to systematically sequence many people to find genetic variants and build a catalogue which we can use as a basis for investigations into disease genetics and which variants may be functional," said Dr Richard Durbin of the Wellcome Trust Sanger Institute in Cambridge and co-chair of the 1,000 Genomes Project.

The project has reached a level of detail unmatched by previous methods and located differences that occur in only 1% of the population. In the pilot phase, 179 people had their whole genomes sequenced and 697 people just the protein-coding regions. In addition, the researchers examined in detail the genomes of two nuclear families. They found that each child had around 60 mutations in its genome that did not exist in either parent.

Evan Eichler of the University of Washington in Seattle said that the project also provides valuable insights into the evolution of our species. "When we compare these 159 humans we've analysed to the great apes, we have the ability to identify the genes and gene families which have expanded specifically in our lineage of evolution since we separated from that of chimpanzees and gorillas," he said. "We find a tantalising set of genes that are important for neural development in terms of neuronal migration and we want to focus on these as helping to define some aspect of the human condition."

The researchers catalogued more than 15m single DNA letter changes in their sample. "Over half those differences haven't been seen before, and these have provided a more complete catalogue of variation than was available previously," said Durbin. "In one person's genome, among the 3m variants that individual will have, over 95% will be in our catalogue." This compares with a catalogue containing 5% of all human variations from the Human Genome Project in 2000 and around 40%-50% five years ago.

In addition to 8m previously unknown variants that affect single nucleotides, the researchers also found 1m structural variants, such as repeated or deleted DNA sequences. "We can see that each individual is carrying a significant number of deleterious mutations, maybe 250 or 300 genes that have defective copies," said Durbin. "We can also look at the effect of recent evolution on human genomes around genes and between populations."

Previous work to build the catalogue of differences in human genes involved a technique called the genome-wide association study. Here, scientists look at small DNA samples from thousands of patients for every disease, comparing them with thousands of control samples from healthy volunteers and looking at hundreds of thousands of genetic differences in each sample.

In June 2007, a British-led collaboration of more than 50 research groups around the world used this technique to examine 17,000 people, identifying a total of 24 new genes linked to bipolar disorder, Crohn's disease, heart disease, type 1 and type 2 diabetes, rheumatoid arthritis and high blood pressure – tripling the number of genes already associated with these conditions. Despite their utility, however, genome-wide association studies can only identify variations that occur in 5%-10% of the population.

In the next phase of the 1,000 Genomes Project, 2,000 samples from 27 populations around the world will be studied over the next two years. David Altshuler of Massachusetts General Hospital and a co-author of the Nature paper said that, with successive phases and more sequences, the catalogue of human genetic variation, which currently contains 95% of the possible differences, will improve. "We see these numbers going to 98%. By the time the 1,000 Genome Project is done, each person who has their genome sequenced, greater than 95% – maybe even 98%-99% – of the variation in that person would already be in the database and could be referenced back. Around 1%-2% of the variation would be unique to that individual and not in that database."

The genome in numbers

3bn Number of DNA letters in the human genome (200 volumes the size of a Manhattan telephone book, which has around 1,000 pages)

20,000-25,000 Number of genes in the genome (though not all scientists agree)

2000 Year the first draft of the human genome was announced to much fanfare at the Clinton White House

2003 Final draft completed to 99.99% accuracy

2500 Number of people whose genomes the 1,000 Genomes Project hopes to sequence

15m Number of single-letter changes identified in the pilot phase

1m Number of small insertions and deletions identified in the pilot phase

4.9 trillion Number of letters of data sequenced by the 1,000 Genomes Project so far