Decoding the first complete human genome, that
of sequencing pioneer Craig Venter, is estimated
to have cost around $10m. The second, that of
James Watson, Nobel laureate and co-discoverer
of DNA’s structure, cost around a tenth of that,
and by June 2009, current market leader Illumina
offered a full genome sequencing service for just
$48,000. Meanwhile, pundits, like the genome
pioneer George Church of Harvard University,
US, predict that we will soon see the price drop
to $1000/genome, and that at this point, genome
sequencing will be as revolutionary as the
introduction of affordable desktop PCs. Complete
Genomics, where Church is a member of the
scientific advisory board, recently announced that
it had sequenced 14 human genomes this year for
as little as $5000/genome.
The sequencing of all 3bn base pairs of the
human genome should not to be confused with
services scanning the genome for common
variants – single nucleotide polymorphisms (SNPs)
– which are already available for less than $100
from companies like 23andme and Knome. Initially
hyped as a panacea for pharmacogenomics,
this type of analysis has revealed very little
information of medical interest. Experts now
agree that complete genomes, and lots of them,
will be needed to understand why individuals
are susceptible to certain diseases and respond
differently to some drugs.
But why is the cost and time required to
sequence a complete genome falling so fast and
dramatically? The current ‘second generation’ of
sequencing machines typically use DNA synthesis
with fluorescent bases, so the machine ‘watches’
as new building blocks are added to a growing
strand and notes the colour of each added unit.
Unlike the living cell, which only needs one double
helix to read out the information stored in it,
sequencing still relies on a multitude of identical
DNA strands.
So the first big improvement sequencers need
is to mimic nature and read the sequence from
a single molecule. Several companies are now
developing single molecule sequencing methods.
The first one to reach the market is Helicos’s
HeliScope system, which is based on sequencing
technology invented by Stanford University, US,
engineer Stephen Quake. Helicos vp and chief
scientific officer, Patrice Milos, says: ‘Single
molecule sequencing will indeed deliver. Over the
next two to four years, we at Helicos indeed see a
path to a complete genome sequence, at 20 times
coverage using the HeliScope, that will be well
below the $1000 threshold.’
In August, Quake became the eighth person to
have his genome published (Nature Biotechnology
2009, 27, 847), an announcement that garnered
much media attention as only three names
appeared on the paper. While this is the first
complete human genome sequenced by single
molecule technology, commentators around the
world quickly jumped on claims of reduced cost.
At just under $50,000, Quake’s sequence is no
cheaper than the commercial offer available from
Illumina, using established technology including
amplification. Genomics blogger Daniel MacArthur
commented: ‘It appears that formal entry into the
human genome sequencing race merely requires
generating a genome sequence of the standard that
second-generation sequencers were achieving in
early 2008, at the same price that they’re charging
right now. That’s a fairly uninspiring goal.’
Meanwhile, other biotech companies are
also developing single molecule technologies.
Pacific Biosciences (PacBio) based at Menlo Park,
California, is developing what they call SMRT
(single molecule real time) DNA sequencing, a
technique based on DNA synthesis taking place in
a tiny hole.
Two novel tricks enable SMRT to sequence
a single strand of DNA. As the strand grows, the
nucleotides’ fluorescent labels, which are attached
to the phosphate groups, are cleaved off. This
reduces background ‘noise’ to produce a cleaner
signal, so that fluorescent markers from a single
DNA strand can be successfully analysed. And to
ensure that only one strand is growing at a time,
the well is so small that only a few strands of DNA
can enter. Diffraction effects within the tiny hole
increase SMRT’s sensitivity by making it unlikely
that more that one fluorescent building block
will be present at any one time. Unlike Helicos’s
technology, which has been criticised for extremely
short reads, typically around 32 bases, SMRT
sequencing is designed to produce long reads and
good processivity.
PacBio’s ceo Hugh Martin says: ‘When we talk
to customers, what they are excited about with the
next generation of DNA sequencing technology
is… the availability of very long, high quality reads
delivered in a very short amount of time. They
are completely agnostic as to how you get those
properties.’
Following the splash caused by Quake’s
genome, PacBio responded swiftly, announcing that
it had acquired $68m in additional funding, raising
total funding to $188m. New investors include
the Wellcome Trust and Monsanto. The company
plans to release a commercial version of its single
molecule sequencer in the second half of 2010.
John Milton, who played a key role in the
development of the Solexa technique which was
bought up by Illumina and is currently leading the
market, is now vp of Research at Oxford Nanopore
Technologies (ONT), a spin-out based on the
nanopore research of Hagan Bayley at Oxford
University, UK. He is leading the team developing
a new single molecule sequencing technology and
believes that they must go one step further than
Helicos and PacBio.
‘The current [second] generation of sequencers
requires amplification and light detection,’ Milton
explains. ‘Then the next step is to use single
molecules and get rid of the [polymerase chain
reaction] PCR. But the ultimate goal must be to
remove the light as well and to develop a technology
that analyses single molecules without optics.’
Using electronic detection instead of optical
devices, Milton says, the sequencers can be
simplified much further, and could become truly
affordable to medical practitioners and research
labs alike.
ONT’s idea is to have a modified version of a
natural membrane pore sitting in a membrane
with an electric potential driving ions through the
channel. Attached to the entrance to the pore sits
an exonuclease, an enzyme that chops bases off the
end of a DNA strand one-by-one. The electric field
pulls the nucleotide released by the enzyme through
the pore and when it passes the narrowest point
– an artificial bottleneck created by incorporation
of a cyclodextrin derivative – it blocks the current
passing through the pore, creating a characteristic
dip in the recorded current trace. Using this
technique ONT researchers can not only distinguish
between the standard four nucleotides, but can
also identify methylated bases, which is important
for an improved understanding of epigenetics – the
regulation of gene activity. ‘We’re on the cusp of a
breakthrough here, which allows digital electronic
readout of sequences, and whose performance will
start where improvements to current technology
level off,’ says ONT’s ceo Gordon Sanghera.
While the Helicos sequencing system is still the
size of a large fridge and comes with an inbuilt
granite slab to stabilise the laser microscopes,
ONT is confident that its technology, which will
be marketed by current leader Illumina, will
ultimately fit into a small desktop box that will be
user-friendly and could become a standard tool in
many hospitals. For instance, the company points
out that oncologists have recently been sequencing
patient’s tumour genomes and comparing them
with healthy cells in order to spot the tumour’s
Achilles’ heel.
So how long before general practitioners are
sequencing a patient’s genome in their surgery or
sequencers are found in hobby labs? ONT does not
want to reveal time plans or prices yet, but it quotes
Illumina president, Jay Flatley, who said earlier this
year that babies born from 2019 onwards will
routinely have their genomes sequenced at birth.
‘We will reach a tipping point as with PCs and the
web,’ Sanghera says.
Michael Gross is a science writer based in
Oxford, UK.