鈥淲E WOULDN鈥橳 have a genome at all if we hadn鈥檛 carried on with the public
effort,鈥� says John Sulston, the former head of the Sanger Centre in
Cambridge.
In 1998, six years after the Human Genome Project began, the effort to
sequence the genome turned into a race when Craig Venter announced that his
company, Celera, would finish it by 2001 using the 鈥渨hole genome shotgun鈥�
method. That led some people to question whether the public effort was a waste
of taxpayers鈥� money. Now Celera is claiming victory. 鈥淭he whole genome assembly
has been an unqualified success,鈥� Eugene Myers of Celera said on Monday.
But the public project was essential to prevent the genome being
鈥減rivatised鈥�, Sulston says. He also rejects the claim that Celera won the race.
鈥淣ow that the papers from the two teams can be inspected, we can for the first
time judge the relative strengths of the two versions,鈥� he says. 鈥淭o get their
sequence, they relied very heavily on the work we put in.鈥� Over half the data in
Celera鈥檚 final version came from the Human Genome Project, Sulston says
(see graphic).
Advertisement
Despite this, Sulston thinks Celera鈥檚 version is not greatly superior to the
HGP鈥檚 draft version. 鈥淩emarkably, his product is very similar to ours. I would
have thought it would be much better. Far from 鈥榳inning the race鈥�, as they have
claimed, their methodology has been found wanting.鈥�
The HGP鈥檚 method involves mapping the genome, by identifying 鈥渟ignposts鈥�
along each chromosome, and then sequencing it region by region. Venter claimed
that the pure shotgun method, in which the entire genome is randomly cut into
pieces, sequenced and then assembled by computers, was much faster and more
efficient.
But Sulston says the version Celera produced using this method is inferior to
the HGP鈥檚 draft. 鈥淚t fails badly both in sequence continuity and long-range
order.鈥� To get its final version, Celera had to resort to a 鈥渃ompartmentalised
shotgun鈥� method. This relies on the public project鈥檚 map information, and is
really a variation on the HGP鈥檚 method, Sulston says. 鈥淲e can say today that we
were right.鈥�
Sulston says that while Celera鈥檚 final version has more gaps, the local
ordering of many of its fragments is more accurate. But this is what you鈥檇
expect, he says, given that Celera had access to the HGP鈥檚 data as well as its
own.
鈥淚t鈥檚 hard to understand where these claims come from,鈥� Venter says. 鈥淭he
whole genome sequencing technique gave a wonderful reconstruction of the
驳别苍辞尘别.鈥�
鈥淲e found that 3 to 5 per cent of the public map was completely
misassembled,鈥� he says. 鈥淲hat adding the [public] data did was make some gaps
smaller and show where some bits were too long. But it didn鈥檛 change the
assembly as a whole.鈥�
Both groups still have a lot of work to do, however. Only 33 per cent of the
HGP鈥檚 sequence has been finalised. It will take up to two years to finish. 鈥淥urs
is not a draft,鈥� Venter insists. 鈥淚t has some gaps but it鈥檚 a highly accurate
sequence with all the pieces in the right order.鈥�
鈥淚t鈥檚 very plainly a draft,鈥� retorts Sulston. 鈥淚t鈥檚 got a very large number
of gaps.鈥�
Controversy also surrounds the decision by the journal Science to
publish Celera鈥檚 papers on the genome even though researchers won鈥檛 be given
unrestricted access to its sequence. 鈥淚f you publish a paper that depends on a
DNA sequence, you put it into one of the public databases,鈥� Sulston says. This
is why the HGP published its papers in a separate journal, Nature,
rather than jointly with Celera, he says.
Venter denies access is restricted. 鈥淥urs is accessible, but through
subscription. From 10 am on Monday, it [has been] available free of charge on
the Internet to any academic for any science they want to do. The one
restriction is that they can鈥檛 take the data we鈥檝e produced and give it to a
commercial competitor that wants to sell it.鈥�
However, among various other restrictions, researchers will only be allowed
to download one million bases a week. 鈥淚 hope this unusual thing, where a major
journal has downgraded publication rules to accommodate the whim of a particular
company, does not create a precedent for the future,鈥� says Martin Bobrow, a
geneticist at Cambridge University.