A special report on the human genomeJune 19th 2010

2.0Biology

Biology2.0.indd 1 08/06/2010 17:11

The Economist June 19th 2010 A special report on the human genome 1

A decade after the human­genome project, writes Geo�rey Carr,biological science is poised on the edge of something wonderful

and lots of money would be made.And then it all went terribly quiet. The

drugs did not appear. Nor did personalisedmedicine. Neither did the genetic under­class. And the money certainly did not ma­terialise. Biotech �rms proved to be just asgood at consuming cash as dotcom start­ups, and with as little return. The casualobserver, then, might be forgiven for think­ing the whole thing a damp squib, and the$3 billion spent on the project to be somuch wasted money. But the casual ob­server would be wrong. As The Economistobserved at the time, the race Dr Venterand Dr Collins had been engaged in was arace not to the �nish but to the starting line.Moreover, compared with the sprint theyhad been running in the closing years ofthe 1990s, the new race marked by thatstarting line was a marathon.

The new race has been dogged by di�­culties from the beginning. There was afalse start (the announcement at the WhiteHouse that the sequence was complete re­lied on a generous de�nition of that word:a truly complete sequence was not pub­lished until 2003). The competitors thenran into numerous obstacles that naturehad strewn on the course. They found at�rst that there were far fewer genes thanthey had expected, only to discover laterthat there were far more. These discoveries

Biology 2.0

TEN years ago, on June 26th 2000, a raceended. The result was declared a dead

heat and both runners won the prize ofshaking the hand of America’s then presi­dent, Bill Clinton, at the White House. Therunners were J. Craig Venter for the privatesector and Francis Collins for the public.The race was to sequence the human ge­nome, all 3 billion genetic letters of it, andthus�as headline writers put it�read thebook of life.

It quite caught the public imagination atthe time. There was the drama of a maver­ick upstart, in the form of Dr Venter and hisnewly created �rm, Celera, taking on themedical establishment, in the form of DrCollins’s International Human GenomeSequencing Consortium. There was thepromise of a cornucopia of new drugs asgenetic targets previously unknown to bi­ologists succumbed to pharmacological in­vestigation. There was talk of an era of�personalised medicine� in which treat­ments would be tailored to an individual’sgenetic make­up. There was the frisson offear that a genetic helotry would becreated, doomed by its DNA to second­class health care, education and employ­ment. And there was, in some quarters, ahope that a biotech boom based on geno­mics might pick up the baton that the inter­net boom had just dropped, and that lots

An audio interview with the author is at

Economist.com/audiovideo/specialreports

A list of sources is at

Economist.com/specialreports

Marathon manGenomics has not yet delivered the drugs,but it will. Page 3

Where are they now?The big beasts of genomics. Page 4

It’s personalIndividualised genomics has yet to take o�.Page 7

The dragon’s DNAThe next advances in genomics may happenin China. Page 8

Inhuman genomesEvery genome on the planet is now up forgrabs, including those that do not yet exist.Page 9

The soul of an old machineGenomics is raising a mirror to humanity,producing some surprising re�ections. Page 11

No hiding placeEveryday genomics is coming, ready or not.Page 13

Also in this section

AcknowledgmentsIn addition to those mentioned in the text, the authorwould like to give special thanks to: David Altshuler, InêsBarroso, Peter Donnelly, Michael Donoghue, Eric Green,Tim Hubbard, Lin Liu, Thomas Lynch, Geo�rey Miller,Thomas Pollard, James Rothman, Ed Scolnick, Frank Slack,Matt State, Michael Stratton, Elia Stupka and ChrisTyler­Smith.

1

2 A special report on the human genome The Economist June 19th 2010

2

1

changed the meaning of the word �gene�.They found the way genes are switched onand o� is at least as important, both biolog­ically and medically, as the composition ofthose genes. They found that their meth­ods for linking genetic variation to diseasewere inadequate. And they found, aboveall, that they did not have enough genomesto work on. Each human genome is di�er­ent, and that matters.

All is revealedOne by one, however, these obstacles arefalling away. As they do so, the science ofbiology is being transformed. It seemsquite likely that future historians of sci­ence will divide biology into the pre­ andpost­genomic eras.

In one way, post­genomic biology�bi­ology 2.0, if you like�has �nally killed theidea of vitalism, the persistent belief thatto explain how living things work, some­thing more is needed than just an under­standing of their physics and chemistry.True, no biologist has really believed in vi­talism for more than a century. Neverthe­less, the promise of genomics, that theparts list of a cell and, by extension, of a liv­ing organism, is �nite and cataloguable,leaves no room for ghosts in the machine.

Viewed another way, though, biology2.0 is actually neo­vitalistic. No one thinksthat a computer is anything more than thesum of its continually changing physicalstates, yet those states can be abstractedinto concepts and processed by a branch oflearning that has come to be known as in­formation science, independently of theshifting pattern of electrical charges insidethe computer’s processor.

So it is with the new biology. The chem­icals in a cell are the hardware. The infor­mation encoded in the DNA is the preload­ed software. The interactions between thecellular chemicals are like the constantlychanging states of processing and memorychips. Though understanding the genomehas proved more complicated than expect­ed, no discovery made so far suggests any­thing other than that all the informationneeded to make a cell is squirreled away inthe DNA. Yet the whole is somehow great­er than the sum of its parts.

Whether the new biology is viewed asrigorously mechanistic or neo­vitalistic,what has become apparent over the pastdecade is that the process by which the ge­nome regulates itself, both directly by onegene telling another what to do and indi­rectly by manipulating the other mole­cules in a cell, is vastly more complicatedand sophisticated than anybody expected.

sort that have given science the Neander­thal genome have been as important to thedevelopment of genomics as intellectualinsights have been. The telescope revolu­tionised astronomy; the microscope, biolo­gy; and the spectroscope, chemistry. Thegenomic revolution depends on two tech­nological changes. One, in computingpower, is generic�though computer­mak­ers are slavering at the amount of data thatbiology 2.0 will need to process, and theamount of kit that will be needed to do theprocessing. This torrent of data, however,is the result of the second technologicalchange that is driving genomics, in thepower of DNA sequencing.

The new lawComputing has, famously, increased in po­tency according to Moore’s law. This saysthat computers double in power roughlyevery two years�an increase of more than30 times over the course of a decade, withconcomitant reductions in cost.

There is, as yet, no sobriquet for its ge­nomic equivalent, but there should be. EricLander, the head of the Broad Institute, inCambridge, Massachusetts, which isAmerica’s largest DNA­sequencing centre,calculates that the cost of DNA sequencingat the institute has fallen to a hundred­thousandth of what it was a decade ago(see chart 1). The genome sequenced by theInternational Human Genome Sequenc­ing Consortium (actually a compositefrom several individuals) took 13 years andcost $3 billion. Now, using the latest se­quencers from Illumina, of San Diego, Cal­ifornia, a human genome can be read ineight days at a cost of about $10,000. Nor isthat the end of the story. Another Califor­nian �rm, Paci�c Biosciences, of MenloPark, has a technology that can read ge­nomes from single DNA molecules. Itthinks that in three years’ time this will beable to map a human genome in 15 minutesfor less than $1,000. And a rival technologybeing developed in Britain by Oxford Na­nopore Technologies aspires to similarspeeds and cost.

This increase in speed and reduction incost is turning the business of biology up­side down. Up until now, �rms that claimto read individual genomes (see box in thenext article) have been using a shortcut.They have employed arrays of DNA

probes, known as gene chips, to look forpre­identi�ed variations in their clients’DNA. Those variations have been discov­ered by scienti�c collaborations such asthe International HapMap Project, whichsearch for mutations of the genetic code

$/MQ20

1Baseline information

Source: Broad Institute

Cost of genome sequencing compared withMoore’s law for computers

Log scale

0.1

1.0

10

100

1,000

10,000

100,000

1999 2002 04 06 08 10

Cost of computing(Moore’s law)

$ per million DNA bases

Yet it now looks tractable in a way that 20years ago it did not. Just as a team of engi­neers, given a rival’s computer, could stripit down and understand it perfectly, so bi­ologists now believe that, in the fullness oftime, they will be able to understand per­fectly how a cell works.

And if cells can be understood com­pletely in this way, then ultimately itshould be possible to understand assem­blages of cells such as animals and plantswith equal completeness. That is a muchmore complicated problem, but it is di�er­ent only in degree, not kind. Moreover, un­derstanding�complete or partial�bringsthe possibility of manipulation. The pastfew weeks have seen an announcementthat may, in retrospect, turn out to havebeen as portentous as the sequencing ofthe human genome: Dr Venter’s construc­tion of an organism with a completely syn­thetic genome. The ability to write new ge­nomes in this way brings true biologicalengineering�as opposed to the tinkeringthat passes for biotechnology at the mo­ment�a step closer.

A second portentous announcement,of the genome of mankind’s closest�albe­it extinct�relative, Neanderthal man,shows the power of biology 2.0 in a di�er­ent way. Putting together some 1.3 billionfragments of 40,000­year­old DNA, con­taminated as they were with the fungi andbacteria of millennia of decay and the per­sonal genetic imprints of the dozens of ar­chaeologists who had handled the bones,demonstrates how far the technology ofgenomics has advanced over the course ofthe past decade. It also shows that biology2.0 can solve the other great question be­sides how life works: how it has evolvedand diversi�ed over the course of time.

As is often the way with scienti�c dis­covery, technological breakthroughs of the

The Economist June 19th 2010 A special report on the human genome 3

2

1

called single­nucleotide polymorphisms,or SNPs, in blocks of DNA called haplo­types. A SNP (pronounced �snip�) is aplace where a lone genetic letter variesfrom person to person. Some 10m SNPs arenow known, but in the forest of 3 billiongenetic letters there is reason to believethey are but a smattering of the total varia­tion. Proper sequencing will reveal the lot.

Finding the sequence�even the fullrange of sequences�is, though, just the be­ginning. You then have to do somethinguseful with the result. This is where thecomputing comes in. Computers allow in­dividual genomes�all 3 billion base pairsof them�to be compared. And not onlyhuman genomes. Cross­species compari­sons are enormously valuable. Laboratoryexperiments on creatures ranging from

yeast to mice can reveal the functions ofgenes in these species. Computer compari­son then shows which human genes corre­spond in DNA sequence and thus, presum­ably, in function, to the genes in these�model� organisms.

Cross­species comparison also showshow species di�er, and thus how theyhave diverged. Comparing DNA from pop­ulations within a species can show howthat species is evolving. Comparing DNA

from individuals within a population canexplain why those individuals di�er fromone another. And comparing the DNA

from cells within an individual can showhow tissues develop and become di�eren­tiated from one another, and what goeswrong in diseases like cancer.

Even before cheap sequencing became

available, huge databases were being builtup. In alliance with pathology samples,doctors’ notes and�most valuable of all�long­term studies of particular groups ofindividuals, genetic information can belinked to what biologists refer to as thephenotype. This is an organism’s outwardexpression: its anatomy, physiology andbehaviour, whether healthy or pathologi­cal. The goal of the new biology is to tiethese things together reliably and to under­stand how the phenotype emerges fromthe genotype.

That will lead to better medical diagno­sis and treatment. It will result in the abilityto manipulate animals, plants, fungi andbacteria to human ends. It will explain thehistory of life. And it will reveal, in pitilessdetail, exactly what it is to be human. 7

�WHERE’S the beef?� is always a rea­sonable question to ask. For the

human genome it can be rephrased slight­ly as �where are the drugs?� It is a questionthat does not exactly make genomicistssquirm, but it puts them on the defensive.

By now, if you had believed the morebullish pronouncements made at the timethe human­genome project was coming tofruition, the pipelines of pharmaceuticalcompanies would have been burstingwith aspiring treatments for everythingfrom Alzheimer’s disease to Zollinger­Elli­son syndrome, as the genes involved inthese illnesses were identi�ed and drugmolecules that could correct malfunctionsof those genes were discovered. In fact, thepipelines are empty; company analysts of­ten seem to regard research as a drain onthe balance­sheet, rather than an asset;and drug companies seem to be reinvent­ing themselves as marketing �rms for es­tablished products. The explanation is atoxic mix of science and economics, butthe result is an industry ripe for disruption.

Don’t count your chickensIn 1990, when the human­genome projectbegan, everybody thought they knewwhat a gene was. It was a stretch of DNA

that could be transcribed by an enzymecalled polymerase into a chemically simi­lar molecule known as RNA. The RNA act­ed as a messenger that was itself translated

into protein molecules in sub­cellular fac­tories called ribosomes. The translationcode was a series of three­letter �words�,called codons, each standing for one of the20 amino­acid molecules that form thecomponents of proteins. The codons werewritten in a four­letter alphabet, A, C, G, T,that abbreviated the names of the chemi­cals of which DNA is made.

It was all very neat. Nobel prizes wereawarded in abundance and, except for afew specialised genes whose RNA was di­rectly involved in the protein­manufactur­ing process, it was understood that the�central dogma� of biology (so describedby Francis Crick, co­discoverer of the struc­ture of DNA) was that one gene equals oneRNA messenger molecule equals one pro­tein. Proteins are the workhorses of cells,acting as enzymes, ion channels, signallingmolecules and structural elements. Andsome proteins act as transcription factors,regulating the output of the genes them­selves. The system made perfect sense.There were a few oddities. Most notably,the best estimate for the amount of DNA

that encoded proteins was only 3% of allthe DNA in the genome. In the rush of self­congratulation, however, no one paid toomuch attention to that fact. The non­geneDNA was dismissively labelled �junk�.

Someone should have taken note,though. A sizeable amount of the junk, itturns out, is transcribed into RNA even

Marathon man

Genomics has not yet delivered the drugs, but it will

4 A special report on the human genome The Economist June 19th 2010

2

1

though it does not make proteins. Instead,the RNA itself is busy doing jobs that wereonce thought to be the prerogative of pro­teins: regulating the transcription of othergenes, protecting cells from viral attack andeven keeping control of bits of DNA thatreally are junk (or, more accurately, are par­asitic on the whole genomic apparatus).

No one knows how many �RNA­only�genes there are, but there could well bemore than 100,000 of them. Ten years ago,only a handful were known. By contrast,the current estimate of the number of pro­tein­coding genes in the human genome is23,000. The RNA­only genes are, more­over, medically signi�cant. They keep pop­ping up, for example, in cancers.

The second layer of complexity�notcompletely unexpected, but certainly un­

derappreciated�is called epigenetics. Thisis the process by which DNA is chemicallyaltered by the addition of a methyl group(a carbon atom and three hydrogens) to ge­netic letter C. Epigenesis is yet another wayof regulating transcription (the methyl­ation stops this happening). It is, however,more permanent than the on/o� switch­ing provided by transcription factors andRNA­only genes. Indeed, it is so perma­nent that it can sometimes be passeddown the generations, leading to a lot ofexcitable talk about the inheritance of ac­quired characteristics�normally regardedas a Darwinian no­no.

Such talk is premature. More perma­nent is not the same as indelible, and epi­genetic changes are not passed on inde�­nitely. Nevertheless, they may help

explain patterns of disease such as late­on­set diabetes. This, some researchers hy­pothesise, might be encouraged by chil­dren inheriting epigenetic patternsappropriate to the diets of their parents butinappropriate to the di�erent, more calor­i�c diets those children are enjoyingthanks to the abundance of modern life.

The third layer of complexity is one thatis only now starting to be explored. Biolo­gists and laymen alike think of the genomeas linear. DNA is, indeed, a long­chain mol­ecule. It is so long, though, that if the 3 bil­lion base pairs were linked together andpulled out straight, the result would be ametre in extent. In reality, DNA is twistedand folded up inside the cell nucleus, withthe result that bits of the molecule thatseem far apart on a map are actually next

SCIENCE reporting usually concentrateson the science, not the scientists.

Though the minds and hands behind theresearch are acknowledged, the real storyis the discovery itself and its place in thejigsaw of human understanding. That,and the fact that modern scienti�c investi­gation tends to be a team e�ort, has dimin­ished the cult of the celebrity scientist.The human­genome project was an ex­ception to this rule. It created some scien­ti�c celebrities and also some celebratedrivalries. Ten years on the reader mightwonder what has happened to them.

The loosest cannon of the lot wasprobably James Watson. Dr Watson, co­discoverer with Francis Crick of the dou­ble­helical structure of DNA, was respon­sible for suggesting the human­genomeproject in the �rst place. At the time he washead of America’s National Centre forHuman Genome Research, part of thecountry’s National Institutes of Health(NIH). He fell out with the NIH, however,over the issue of patenting DNA se­quences called expressed sequence tags.He opposed this, arguing that �youshouldn’t patent something a monkeycould do.� That did not endear him toCraig Venter, who had created the DNA

in question. Dr Watson was replaced byFrancis Collins, a man regarded by somebiologists as ideologically unsound be­

cause he is a born­again Christian. Dr Wat­son continued as head of the Cold SpringHarbour genetics laboratory until 2007,when he made some injudicious remarksabout genetics and black people andfound himself suddenly retired.

Dr Venter, too, left the NIH in the wakeof the expressed­sequence­tag incident.At �rst he teamed up with Bill Haseltine,a virus geneticist with a record as an entre­preneur, to start an institute and a com­pany, Human Genome Sciences, to ex­ploit expressed sequence tags. The twofailed to see eye to eye, though, and DrVenter went on to help create a second�rm, Celera, in the hope of beating thepublic project to the human genome. Heused a new DNA­sequencing techniquecalled whole­genome shotgunning thathe and a colleague, Hamilton Smith, hadinvented, and patented a good tranche ofhuman genes on the way. That was anath­ema to Dr Collins and his British counter­part, John Sulston (who was head of theWellcome Trust’s Sanger Centre, nowknown as the Sanger Institute, which didabout a third of the public project), whowanted genes to be public goods andstarted racing Celera to stop the �rm �nd­ing genes �rst.

At this point, heads were knocked to­gether in the public project, chie�y by EricLander, of the Whitehead Institute in

Cambridge, Massachusetts, whose out�tachieved e�ective leadership of theAmerican arm of the project. It was DrLander whose name led the list of re­searchers on the paper eventually pub­lished by Nature in 2001.

Celera’s version of the genome waspublished by Nature’s rival, Science. Un­fortunately, this scienti�c triumph did notproduce much in the way of revenue andDr Venter was sacked from the �rm in2002. (Dr Haseltine left Human GenomeSciences in 2004.) Dr Venter then went ona round­the­world cruise on his yacht, col­lecting bacterial samples from the sea fora project he dubbed the Global OceanSampling Expedition. He also set up yetanother research institute (which un­veiled an organism with an arti�cial ge­nome last month) and another commer­cial arm, called Synthetic Genomics, incollaboration with Dr Smith.

Dr Lander, too, has added to the num­ber of America’s research laboratories.With money from two Californian bene­factors of that name he has set up theBroad Institute, America’s largest ge­nome­sequencing lab, next door to theWhitehead. Dr Sulston, meanwhile, hasstarted a scienti�c­ethics institute at Man­chester University and Dr Collins hasbagged one of the top prizes in Americanscience. He is now head of the NIH.

The big beasts of genomics Where are they now?

The Economist June 19th 2010 A special report on the human genome 5

2

1

to each other in the nucleus. How much this matters is almost com­

pletely obscure. What is known is thatthere are often active zones of DNA tran­scription within a nucleus that seem to bemuch bigger than the width of a strand ofDNA and its associated proteins. This sug­gests that genes apparently a long wayfrom one another are actually, in somesense, collaborating.

All this biological complexity would bebad enough by itself for drugmakers seek­ing a quiet life. The other problem, though,was a quite monumental naivety aboutthe ease of linking newly discovered genesto diseases and disease processes. This hasactually proved �endishly di�cult.

Lighten our darknessThe favoured approach has been the ge­nome­wide association study, or GWAS.Hundreds of these have been carried outover the past �ve years or so. The ideasounds sensible: gather samples from peo­ple with and without particular diseasesand look for associations between thosediseases and particular genetic mutations,in the form of SNPs. The practice, however,has not really come up with the goods.

The thinking behind GWAS was that itwould expose multigenic diseases. Theseare conditions that seem to run in familiesbut do not obey the clear­cut laws of inher­itance laid down in the 19th century byGregor Mendel. Those diseases that do be­have in a Mendelian way�haemophiliaand sickle­cell anaemia, for example�areclosely tied to the mutational failure of in­dividual genes (a blood­clotting factor andone of the genes for haemoglobin, respec­tively, for these two diseases). The tenden­cy of people in some families to su�erheart disease, strokes, late­onset diabetes,Alzheimer’s disease and so on is, by con­trast, less clear­cut. Environmental factorsare obviously involved. But mutations are,too�just not single, large­e�ect mutationslike those that cause haemophilia and sick­le­cell anaemia. Instead, the pattern of in­heritance suggests that many mutations ofsmall individual e�ect come together toproduce a risk rather than a certainty.

GWAS has not been a total failure. Ithas revealed lots of mutations of small ef­fect. On average, though, these add up toonly 10% of the total heritability of any giv­en disease. Mendelian e�ects add aboutanother 1%. The rest, in a phrase that geneti­cists have borrowed from physicists, is re­ferred to as �dark matter�. These mutationsappear to be tremendously important, yetneither Mendelian nor GWAS techniques

can detect them. Mendelian mutations arenoticed because they are rare and power­ful. GWAS mutations are seen because,though puny, they are common. The darkmatter lies in the middle: too rare forGWAS but not powerful enough to leave aclear Mendelian signal. Bigger GWAS,with more statistical power, may help a bit,but clearly new methods are needed. Onewill be to deploy whole­genome sequenc­ing more widely, now that it is becomingso much cheaper. And here the study ofone particular sort of disease, cancer, isleading the way.

Compare and contrastCancer is at the vanguard of genomicmedicine for two reasons. One is that on­cologists and their patients (and also theregulators of medical practice) are oftenwilling to take risks that would be unac­ceptable if the alternative were not a horri­ble death. The other is that cancer is nowknown unequivocally to be a genetic dis­ease. Its environmental correlates (smok­ing, for example) act not by poisoning cellsdirectly but by promoting mutations inthose cells’ DNA. Such somatic mutations,as those in body cells are known, can causechaos in an individual’s organs, but are notpassed to his or her o�spring.

In the case of cancer, an accumulationof somatic mutations causes a breakdownof the regulatory mechanisms that stop acell from multiplying uncontrollably. Withthe brakes o�, the cycle of division, growthand further division continues unabateduntil the body can no longer support bothhealthy tissue and tumour.

One lesson that genomics taught oncol­ogy early on is that cancers which looksimilar under the microscope can havecompletely di�erent genetic causes and

thus require di�erent treatments. That gen­eral observation should soon be rein­forced in detail by a project run by the In­ternational Cancer Genome Consortium(ICGC), a collaboration of researchers in 11countries. The plan is to take advantage ofthe falling cost of sequencing to collect fullDNA sequences from 500 people su�eringfrom each of 50 types of cancer. Not onlywill the cancerous tissue be sampled, sowill healthy tissue from each patient.

Comparing the healthy and the cancer­ous tissue in each individual will revealthe somatic mutations which that individ­ual has undergone. Comparing canceroustissue from di�erent individuals will showwhich mutations are important. This isnecessary because in cancer patients thegenes which control the proofreading ofnew DNA strands often become damaged,so that mutations accumulate much faster.That means crucial mutations are morelikely to happen, but also that in any givencancer there is a lot of mutational �noise�.

Until now, this has made it more di�­cult to discover which mutations are im­portant and which merely incidental. Theresult of the ICGC study should be a near­complete understanding of cancer at thegenetic level. That will help diagnosis andtreatment (allowing doctors to choose ap­propriate drugs the �rst time round, ratherthan employing trial and error) and, withluck, should promote the development ofnew treatments.

But identifying the dodgy genes is onlythe �rst step to such treatments. Not allgene products are, in the argot, �drugable�.And this is where the economics comes in.

Todd Golub of the Broad Institute, inCambridge, Massachusetts, reckons drug�rms have got rather lazy about pursuingleads. For example, many oncogenes, as

6 A special report on the human genome The Economist June 19th 2010

2

1

those whose breakage causes cancer areknown, encode proteins called kinases.These are enzymes which are involved inintracellular signalling pathways. A luckybreak some years ago revealed a systemat­ic way of attacking kinases with small mol­ecules that block their activation. Re­searchers with putative anti­kinase drugsare thus welcomed by venture­capital�rms. The odds of success are understoodand the time to market is tolerable. That isin marked contrast to, say, drugs that mightcontrol transcription factors. A failed tran­scription­factor gene is as common a causeof cancer as a failed kinase gene. Transcrip­tion factors, though, are not regarded asdrugable. No systematic way of dealingwith them has yet been discovered.

That is not the venture capitalists’ fault.Is it the drug companies’ fault? They mightargue that they are not in the business ofbasic research. On the other hand, a break­through in this area would create a wholenew line of business. However, if thatbreakthrough were a conceptual one thatcould not be protected by patent ratherthan, say, an individual molecule thatcould be patented, then other �rms wouldbe able to freeride on the discoverer’s ex­pensive research.

Fair sharesDr Golub has a suggestion to break the im­passe. Independent laboratories like theBroad could act as honest brokers for gen­eral research paid for by a cabal of all thebig drug companies. Having paid equally,all would bene�t equally. This being basicresearch, openly published, such collabo­

ration would probably be permitted byantitrust laws. Something similar was triedat the beginning of SNP studies (though ad­mittedly those have not yet led to much inthe way of medicine). At the moment thedrug �rms do not seem interested. Perhapsthat will change as their pipelines empty.

Despite such obstacles, genomics hasalready led to some successes in cancertreatment. The astonishing possibilitiescan be seen in the two photographs on thispage. They are of the same individual be­fore and after treatment with a moleculecode­named PLX4032. The shadows are tu­mours from secondary melanoma, one ofthe most aggressive cancers known.PLX4032 cleared them almost completely.It was designed speci�cally to interact withthe protein produced by a particular mu­tated version of a gene called B­RAF. Thismutation, called V600E, has been found tobe involved in 60% of cases of malignantmelanoma and, less commonly, in othercancers. PLX4032 inhibits the activity ofthe mutated protein and causes cells con­taining it to die.

In this case, the system has worked as itis supposed to. The protein encoded by B­

RAF is a kinase (and therefore familiar toventure capitalists). The initial develop­ment was done by a small biotech �rmcalled Plexxikon, co­founded by JosephSchlessinger of Yale University, one of theearly researchers on B­RAF. The moleculehas now been picked up by a big drug com­pany, Roche, which is paying for phase IIItrials, the last stage before a drug is o�eredto the authorities for approval. If all goeswell PLX4032, no doubt sporting a more

friendly name, will soon be available forthose su�ering from melanoma, and willalso be undergoing trials in other sorts oftumour in which V600E is implicated

There is a sting in the tail. For the mo­ment the protective e�ects of PLX4032 lastonly for six months or so. Presumably, fur­ther mutations bypass the V600E�precise­ly the sort of question that the ICGC pro­ject is designed to address. Once thosemutations are identi�ed, the hope is thatdrugs against them can be developed, too.If that proves possible, all of the pathwaysthat lead to cancer could be blocked. Thatwould, in e�ect, be a cure.

As a demonstration of what genomicscan do, PLX4032 is impressive. The ques­tion is, can this sort of thing be done withother sorts of disease? One of Dr Schlessin­ger’s colleagues at Yale, Richard Lifton,thinks it can. He points to a number of re­cently discovered genes that are now thesubject of investigation by drug compa­nies. PCSK9, which encodes an enzyme in­volved in cholesterol metabolism, is a tar­get for the prevention of heart disease.People with mutated versions of ROMK,the gene for type of potassium­ion chan­nel, have abnormally low blood pressureso the search is on for a drug that tweaksthe unmutated version of the channel, tolower the pressure of people with hyper­tension. Those with mutated versions ofSCN9A, which encodes a particular sodi­um­ion channel, are insensitive to pain.Tinkering with this might produce a supe­rior analgesic. And BACE, the gene for anenzyme called beta secretase, is involvedin Alzheimer’s disease. Inhibiting its actionmay delay the progress of that condition.

A little knowledgeThis handful of promising candidates,though, shows up the drug companies’real gripe about genomics. It is one thing to�nd a gene in the genome; it is quite anoth­er to �nd out what it does; and another stillto understand whether that knowledgehas any medical value. Until these pointsare dealt with, the drugmaking machinethat genomics once promised to becomecannot be built.

Thinking on a grander scale is needed.One bold thinker is George Church of theHarvard Medical School. Dr Church’s Per­sonal Genome Project (PGP) proposes tocollect samples and medical data from100,000 people and use the newly emerg­ing mass­sequencing techniques to recordthe entire genomes of each of them. In away, the PGP will be competing with anumber of commercial operations (see

The Economist June 19th 2010 A special report on the human genome 7

2

1

box below). The two di�erences are that itwill be free to enter and that all the infor­mation will be publicly available.

That open access is a bold idea�andPGP is made bolder by the fact that DrChurch hopes to use new techniques toconvert some sample cells into stem cells,from which all of the body’s tissue typescan be grown. This will enable the pro­ject’s researchers to do genetic investiga­tion on a tissue­by­tissue basis.

The United Kingdom’s Biobank is evenmore ambitious. It is a government­runproject that aims to collect tissue samplesand medical data from 500,000 people inBritain. That will allow an analysis of long­term correlations between genes (as wellas lifestyles) and health, though in contrastto the PGP the participants’ anonymitywill be preserved.

�Retro�tting� existing data­collection

projects is also yielding results. As an ex­ample of what can be achieved, Dr Liftonpoints to the Framingham Heart Study.This began in 1948 and has followed sever­al generations in the town (near Boston) totry to disentangle the causes of heart dis­ease and strokes. It has revealed a numberof genes that may prove classic examplesof dark matter.

If present as a single copy inheritedfrom either mother or father, certain muta­tions of these genes protect against heartdisease and strokes. Those mutations arenot, however, common enough to becaught by conventional GWAS. On theother hand, individuals who inherit twocopies of them, one from each parent, willbe stillborn and so they cannot be detectedby classic Mendelian counts either.

This sort of result suggests the logicalthing to do would be to throw everyone’s

medical records into the genetic maw. Thebigger the sample, the more robust theeventual result. The main objection to thatis privacy, and the extent to which this real­ly is an issue may be shown by the successor otherwise of Dr Church’s approach. ThePGP began recruiting in earnest in March,so all should soon be clear. If the project isa success, the view that a person’s DNA ishis own business may fade away.

Even when the data from the biobanksand the personal genome projects are in,though, they will still have to be turnedinto knowledge that can be converted intopharmaceuticals. Turning data into knowl­edge is the job of people like Aviv Regev ofthe Broad Institute. It will be a hard slog, in­volving yet more sequencing (of RNA,rather than DNA, to see which genes areactually active in particular sorts of cell)and lots of computing (to show up the cor­

ONE way of trying to make money outof the new genomic knowledge has

been to o�er what has come to be knownas �personal genomics�. The results, to putit charitably, have been mixed, and forgood reason. The price point is wrong, ob­serves Douglas Fambrough of Oxford Bio­science Partners, a venture­capital �rmbased in Boston. What you learn fromlooking at your genome is not yet worththe price you have to pay. Either the pricemust come down or the value of the pro­duct must rise. Both may happen whenthe latest generation of DNA sequencersare more widely deployed, but at the mo­ment most personal­sequencing compa­nies use gene chips to give a SNP pro�le,rather than o�ering a complete sequence.

Two of the earliest entrants to the �eldwere deCODE and 23andMe. DeCODE, anIcelandic �rm whose aspirations to be­come a full­�edged pharmaceutical com­pany were dealt a blow when it wentthrough a bankruptcy restructuring earli­er this year, charges $2,000 to search asample for 1m SNPs predictive of 50 genet­ic traits, not all of them diseases. Theragenmakes a similar o�er from South Korea.23andMe, based in Mountain View, Cali­fornia, charges $499 to search more thanhalf a million SNPs for signs of 154 traits.

Navigenics, down the road in Foster City,restricts its analysis ($999) to 28 healthconditions and 12 drug responses �thatyou and your doctor can act on�. Com­plete Genomics, another Californian �rm(Mountain View again), plans to leapfrogthe chip­based crowd by o�ering custom­ers full DNA sequences using a complicat­ed proprietary technology that will not,initially, be for sale to other users. AndKnome, a �rm based in Cambridge, Mas­sachusetts, o�ers a bespoke whole­ge­nome service for the discerning client at$68,500 a pop.

Broadly, personal genomics o�ers twoservices. One is to trace your ancestryback through humanity’s family tree to itsroots in north­east Africa (an o�er like thatof the Genographic Project, described lat­er in this report). Indeed, a similar serviceis also available to pet owners, courtesy ofMars (who make a lot of pet food as wellas confectionery). The other service is pre­dictive medicine�a list of genetic varia­tions that might put you at higher­than­average (or, indeed, lower­than­average)risk of developing particular diseases.

Predictive medicine is a controversialarea. Being told that you have an in­creased chance of illness over the courseof your life can be upsetting, particularly

if no treatment or pre­emptive action ispossible (hence Navigenics’s caveats).Worse, that worry may be misplaced; anincreased chance is not a certainty. Con­versely, it is now clear that the GWAS

studies on which many of the correlationsare based have uncovered only a smallpart of the risk, so not showing up as be­ing in danger does not put someone in theclear. On the bright side, this sort of studycan sometimes reveal the precise natureof a set of symptoms, which might a�ectwhich medicines are used to treat them.

Such precision is one aspect of a �eldcalled pharmacogenomics, which seeksto match drugs to a patient’s genome. Asecond aspect is that genetic knowledgecan sometimes predict adverse reactionsto drugs that have proved safe for somepeople to take but dangerous to others.

Pharmacogenomics, then, is a type ofpredictive medicine that could be a boonfor patient and drug company alike. It al­lows prescriptions to be safer and more ef­fective, and enables �rms that want totake molecules through clinical trials to re­strict the tests to people who are likely torespond well. That makes trials cheaperand more likely to succeed. It also ensuresthat the drug, once approved, is given onlyto those who will bene�t from it.

Individualised genomics has yetto take o�It’s personal

8 A special report on the human genome The Economist June 19th 2010

2

1

IN AN old printing works on an obscureindustrial estate in Hong Kong’s New Ter­

ritories a little bit of history is being made.Most of the �ve­storey building is dustyand derelict. One �oor, however, is state­of­the­art. The paintwork shines. The met­al gleams. And in the largest room the elec­trical sockets in the �oor sit in serried ranksawaiting contact.

That contact will shortly be made withthe delivery of 120 spanking new top­of­the­range Illumina sequencing machines.When they have all been installed thebuilding will, so it is claimed, have more

DNA­sequencing capacity than the wholeof the United States. And that is just thestart. According to Alex Wong, who runsthe facility, the other four �oors will alsosoon be refurbished and the whole build­ing will become a powerhouse ready togenerate information for biology 2.0.

The building belongs to the BGI, onceknown as the Beijing Genomics Institute.Mr Wong manages the institute’s HongKong operation, but the institute itself isbased over the border in the People’s Re­public proper, in Shenzhen. The BGI itselfis one part�arguably the leading one�of

China’s e�ort to show that it can be the sci­enti�c peer of the West.

Its boss, Yang Huangming, is certainlythe peer of people like Dr Venter, Dr Land­er and Dr Collins. He is a man on a missionto make the BGI the �rst global genomicsoperation. Part of the reason for buildinghis newest sequencing centre in HongKong is to reassure researchers from othercountries that the facility will operate in­side a reliable legal framework. If all goeswell, laboratories in North America andEurope will follow.

The BGI began in 1999, when Dr Yangmuscled his way into the human­genomeproject, cornering part of the tip of chro­mosome three (about 1% of the total hu­man genome) as the Chinese contributionto that international project. From thishumble beginning it now plans to se­quence 200 full human genomes as part ofan international collaboration called the1,000­genome project. Half these genomeswill be Chinese, but the institute’s re­searchers intend to sample the full geo­graphical range of humanity. And not onlyhuman genomes. The BGI has alreadysolved the genomes of rice, cucumbers,soyabeans and sorghum, honeybees, wa­ter �eas, pandas, lizards and silkworms,and some 40 other species of plant and an­imal, along with over 1,000 bacteria.

And it, too, is interested in cancer. Ac­cording to Dr Yang the institute will notmerely compare healthy and tumorous tis­sue from the same individuals, as the Inter­national Cancer Genome Consortium (ofwhich it is a part) plans to do, it will actual­ly be able to follow the pattern of muta­tion, in the order that it happened, withinan individual that has led to his cancer.That may allow pre­emptive treatment to

The dragon’s DNA

The next advances in genomics may happen in China

relations in activity which indicate thatparticular molecules are parts of the samebiological pathway).

The way Dr Regev describes it is verymuch like electronics. First, the compo­nents (the biological equivalents of transis­tors, diodes and resistors) must be identi­�ed. That might now be thought of as�classical� genomics. Then those compo­nents need to be assembled into modules(the equivalent of a computer’s logic

gates). That is where the RNA sequencing(along with a host of other tools) comes in.Lastly, the modules can be linked up as cir­cuits and the whole apparatus of the cellshould become clear.

The practical advantage of this knowl­edge will be that the cell’s circuitry can bealtered to bypass broken bits rather than�xing the break itself. That opens up awhole new way of thinking about drug de­velopment. Add that to the plethora of

new targets, in the form not only of the ex­tra protein­coding genes discovered by theoriginal genome project but also of theRNA­only genes and the epigenome, andthe long­term opportunities for pharma­ceutics ought to be bright. Who will takeadvantage of those opportunities, though,remains to be seen. For if it is true that therich world’s established pharmaceuticalcompanies have become stodgy, the geno­mic future may lie elsewhere. 7

The Economist June 19th 2010 A special report on the human genome 9

2

1

IF THE history books do come to recog­nise the idea of biology 2.0, then the date

it began may well be recorded as May 20th2010. That was the day when Craig Venterannounced JCVI­syn1.0, the world’s �rstliving organism with a completely synthet­ic genome.

The Frankencell project, as it wasknown jokingly at the beginning, hadbeen going for 15 years�ever since Dr Ven­ter started to wonder what was the mini­mal genome necessary to support a livingorganism. To �nd out, he took a bacteriumcalled Micoplasma genitalium, which has aparticularly short genome anyway, andknocked its genes out one at a time to seewhich the bug could live without (at leastin the cushy circumstances of a laboratoryPetri dish). The answer was around 100 ofits original complement of 485.

The genetic �exibility this hints at�of acore set of genes and a penumbra of othersuseful in particular circumstances�has,over the past decade, been con�rmed formany other species of bacteria. Indeed,the way biologists think about the wholeidea of �species� when they study thesemicro­organisms is beginning to shift rap­idly. This is part of a general broadening ofgenomics. Though navel­gazing into Homosapiens’s own genome remains of intense

interest, the study and manipulation ofnon­human genomes may ultimatelyhave greater impact.

Dr Venter certainly hopes so. His com­pany, Synthetic Genomics, based in SanDiego, plans to patent the new bug. It ar­gues that although it is a living organism,which would normally be outside thescope of patent law, it is also a true artefact,not just the product of selective breeding.The �rm will then be able to use it, and themethod used to construct it, in its pro­grammes to make fuels and vaccines.

Technology and magicOn the other side of America, in Boston,George Church is taking a di�erent ap­proach. Unlike Dr Venter, who focuses hisenergy on one �rm, Dr Church is a promis­cuous entrepreneur. He has been involvedin the foundation of several companies, in­cluding LS9 and Joule Biotechnologies(which hope to manufacture biofuels) andMicrobia (which plans to make specialitychemicals). Like Dr Venter, Dr Church hassomething up his sleeve. This is MAGE, asomewhat contrived acronym for multi­plex automated genome engineering.

Instead of making new genomes fromscratch, Dr Church plans to make lots ofparallel changes in existing ones. The idea

is to induce simultaneous random muta­tions in all of the genes in a particular cellu­lar pathway by introducing pieces of DNA

which match parts of those genes, butwhich are attached to short sequences thatdo not. As a cell replicates, the foreign DNA

is absorbed and the genes in question aremodi�ed by the non­matching short se­quences. Thousands upon thousands ofdi�erent versions of the pathway are thuscreated, and all are subsequently isolatedand tested to see which are most e�ective.The process is then repeated on the win­ners until the desired outcome is achieved.

Inhuman genomes

Every genome on the planet is now up for grabs, including those that do not yet exist

01 03 05 07 09

2Genes unzipped

Source: GenBank*Animals, plants, fungi and protozoans

†Bacteria and archaea ‡Year so far

Number of fully sequenced species

0

10

20

30

40

0

300

600

900

1,200

2000 02 04 06 08 10‡

Eukaryotes* Prokaryotes†

be developed for people whose tumoursare not yet malignant. Indeed, as the priceof sequencing drops, this �internal phylo­genetics�, as Dr Yang calls it, might be ex­tended to trace the pattern of mutationthat develops in even an apparentlyhealthy body as cells proliferate within it.That may yield nothing interesting. On theother hand it might help explain patternsof disease associated with ageing as cellswhose ancestors were genetically identi­cal slowly diverge from one another.

A better balsaThe BGI also has non­medical ambitions.Its researchers are examining fast­growingplants with interesting structural proper­ties, such as balsa, a lightweight SouthAmerican wood familiar to generations ofschoolboy model­makers, and bamboo, atraditional construction material in China.They are experimenting, too, with animal

cloning. The BGI was the �rst out�t toclone pigs, and it has developed a new andmore e�ective way of cloning mammalsthat might ultimately be applied to hu­mans, if that were ever permitted.

But the organisation is involved in evenmore controversial projects. It is about toembark on a search for the genetic under­pinning of intelligence. Two thousand Chi­nese schoolchildren will have 2,000 oftheir protein­coding genes sampled, andthe results correlated with their test scoresat school. Though it will cover less than atenth of the total number of protein­cod­ing genes, it will be the largest­scale exami­nation to date of the idea that di�erencesbetween individuals’ intelligence scoresare partly due to di�erences in their DNA.

Dr Yang is also candid about the pos­sibility of the 1,000­genome project reveal­ing systematic geographical di�erences inhuman genetics�or, to put it politically in­

correctly, racial di�erences. The di�erencesthat have come to light so far are not in sen­sitive areas such as intelligence. But if hisstudy of schoolchildren does �nd genesthat help control intelligence, a compari­son with the results of the 1,000­genomeproject will be only a mouse­click away.

At the moment this frenetic activity ispaid for mostly by regional developmentgrants and loans from state­owned Chi­nese banks, but Dr Yang hopes to go prop­erly commercial. The Hong Kong opera­tion will work partly as a contractor, andMr Wong hopes to persuade biologistsaround the world to send their samples inand have them sequenced there ratherthan relying on their own universities todo the sequencing. Whether the BGI’s re­searchers can turn their mass­producedDNA sequences into new scienti�c in­sights and bankable products remains tobe seen, but the world is watching. 7

10 A special report on the human genome The Economist June 19th 2010

2

1

This can replicate at a cost of thousands ofdollars the sort of genetic modi�cationsthat have previously cost millions.

Even without the new platforms, non­human genomics is beginning to pay divi­dends. Several �rms, including SyntheticGenomics, LS9 and Joule, are engineeringmicro­organisms (sometimes bacteria,sometimes single­celled algae) to turn outbiofuels resembling the petrol, diesel andkerosene that people put in cars and air­craft. Existing biofuels, based on ethanol,are less good. Ethanol is corrosive and hasless energy per litre than petrol and diesel.

One �rm, Amyris Biotechnologies, is al­ready scaling up to industrial productionof such biofuel, but in Brazil, where cheapcane sugar provides the raw material, rath­er than in the United States, where it isbased. Joule plans to use an even cheaperraw material: the carbon­dioxide exhaustfrom power stations. It is one of the �rmsworking on single­celled algae, tweakingtheir metabolic pathways to improve therate at which CO2 is �xed by photosynthe­sis and then converted into hydrocarbonsthat can be used in cars.

Fuels are an attractive alternative todrugs for the new generation of syntheticbiologists because they are not subject toregulatory whim to the extent that drugsare. If anything, regulation is likely to fa­vour them because their raw material is, ei­ther directly or indirectly, carbon dioxidethat has come from the atmosphere orwould end up there. That makes themgreen in the eyes of governments, andtherefore a good thing.

The smell of moneyFuel, however, is a low­value commodity.A more pro�table way to avoid the regula­tors may be to make complicated high­price chemicals such as fragrances. This iswhat Allylix, of San Diego, California, isdoing. The �rm’s founders realised thatbiological synthesis of certain sorts of mol­ecule is much more e�cient than chemicalsynthesis. Many organic molecules con­tain what are known as chiral centres.These are places where the atoms can bearranged either left­handed or right­hand­ed. In biochemistry, handedness can mat­ter. Left­ and right­handed versions may,for example, smell di�erent. Traditionalchemical synthesis cannot distinguish be­tween left­ and right­handed versions, sothey have to be separated afterwards,which is tedious. Moreover, if there are lotsof chiral centres in a molecule, and eachmatters, the yield of the version with theright combination can be minuscule.

Allylix gets around this by engineeringthe genes for new biological pathways intoyeast cells. The molecular family it concen­trates on is the terpenes, which are used asfragrances and �avours. Some are verycostly. Sandalwood essence, for example,is a terpene, and the demand for its potentsmell means the tree it comes from is be­coming rare. Allylix has duplicated thesmell of sandalwood industrially, by ex­tracting the genes for the relevant enzymesfrom sandalwood trees. Indeed, its re­searchers have improved on nature. Theyhave identi�ed the parts of the enzymemolecules that carry out the reactions, andtinkered directly with the DNA that de­scribes these parts, in order to improvetheir e�ciency. Microbia plans somethingsimilar, using Dr Church’s MAGE technol­ogy, though its �rst products will be colour­ings rather than fragrances.

If genes are to be the raw material of anew technology, then it would be usefulfor researchers to know how many thereare out there. The answer is, a staggeringnumber. Most of them are bacterial.Though the average bacterium has fewerthan 5,000 genes, compared with around20,000 protein­coding genes in the aver­age mammal (bacteria do not go in muchfor RNA­only genes), there are lots of spe­cies of bacteria. In his round­the­worldcruise after he left Celera, Dr Venter reck­oned he identi�ed 5m new bacterialgenes�and that was just a start.

Measuring bacterial diversity in genes

rather than species makes sense because itis no longer obvious exactly what consti­tutes a bacterial species. In the view of Ju­lian Parkhill, of the Sanger Institute, nearCambridge, England, bacteriologists needto shift the focus of their investigationsfrom organisms to systems. The geneticists’workhorse, E. coli, for example, has about4,500 genes. Only 1,500­2,000 of these arealways present, however. The remainderof any given E. coli bacterium’s genome isdrawn from a pool of about 20,000 othergenes that the organisms swap with gayabandon. In only a tenuous sense, then, isE. coli a species in the way that, say hu­mans or mice are species.

The same thing is true of other well­studied bugs, such as those that cause ty­phus and plague, and is likely to be true ofmost bacteria. Thirty­four years ago Rich­ard Dawkins, an evolutionary biologist atOxford University, proposed the idea that�sel�sh genes�, not individual organismsor entire species, are the units on whichevolution acts. In the case of bacteria,which seem to exchange genes promiscu­ously, that seems an excellent way of look­ing at things.

In a sense, such pro�igacy extends tohumanity, too. Add in the genes of the bac­teria that live in peaceable collaborationwith the average human (in his gut, on hisskin and so on), and the �human� genomeexpands from 23,000 protein­coding genesto something more like 3m, according toFrancis Collins. Nor is it pure sophistry tothink of these genes as part of an extendedhuman genome. Many of the bacteria inquestion are genuinely mutualistic withtheir hosts, helping the process of diges­tion or warding o� pathogenic bugs.

This is scarcely explored territory. DrCollins estimates that 90% of these humansymbionts cannot be cultured by normallaboratory methods. America’s NationalInstitutes of Health, which he heads, isbacking yet another of genomics’ big col­laborative e�orts, the Human MicrobiomeProject, to help put that right.

Nor is the human microbiome merelyof academic interest. For example, somethink that the mix of bacteria in a person’sgut can a�ect his chances of becomingobese. If those studying the genetics ofobesity concentrate all their e�orts on thegenes in human cells, they might thus belooking in the wrong place.

The other area where genomics is likelyto have a big practical impact is agriculture.At the moment, despite the brouhaha theyhave created in some countries, geneticallymodi�ed plants are primitive things. Most

The Economist June 19th 2010 A special report on the human genome 11

2

1

of them have had but a single genetweaked, either to make them poisonousto pestilential insects or resistant to a par­ticular herbicide so that it can be used free­ly. Even so, GM crops are big business. A re­cent report by the International Service forthe Acquisition of Agri­biotech Applica­tions, a not­for­pro�t out�t that monitorsthe use of GM crops, suggested that morethan three­quarters of the world’s soya­bean plants are genetically modi�ed,along with half the cotton and more than aquarter of the maize.

Fire burn and cauldron bubbleThe next generation will be bigger busi­ness still. Ceres, a small biotech �rm basedin Thousand Oaks, California, is collabo­rating with Monsanto, a giant agribusinesscompany, to make crops better in all sortsof ways. Their genes are being tweaked toincrease the plants’ drought­resistance andimprove their absorption of nitrogen.

Ceres’s collaboration with Monsantoinvolves traditional crops such as maize,but Ceres is also interested in the energybusiness. Before fossil fuels became ubiq­uitous, plants�in the form of �rewood�were one of humanity’s main sources ofpower. Ceres hopes those days will soonreturn. One way to release useful energyfrom plant matter is to ferment it into bio­fuels, as Synthetic Genomics, LS9 andAmyris are trying to do. Ceres is involved

in this business, too, but Richard Hamilton,the �rm’s boss, is hedging his bets.

Whether biofuels have a big future is amoot point. At the moment the car indus­try seems to view electricity as the motivepower of the future. But that does not wor­ry Dr Hamilton because even if the cars ofthe future are electric, the electricity willhave to come from somewhere�and if thatsomewhere is not fossil fuels, then it mightbe from burning plant matter.

Viewed in this light, plant matter is justan alternative form of solar energy. In hot,dry parts of the world, turning sunlightinto electricity directly with solar cells orindirectly with solar­powered steam tur­bines makes sense. In places where it iscooler and wetter, the equation changes.Growing plants and burning them may bea better way. Moreover, plant matter, oncegrown, is available 24 hours a day. It canthus provide an electrical baseload in away that traditional solar power (whichgoes o� at night) cannot.

To this end, Ceres is tinkering with threespecies of grass: Miscanthus, switchgrassand sorghum. Its researchers have �ddledwith the genes of these so­called energycrops to increase the amount of lignin inthem, at the expense of carbohydrates likecellulose. That makes them more�woody�, increasing their energy content.Ceres is also applying to its energy cropsthe sorts of genetic modi�cation that it has

been developing in collaboration withMonsanto for use in food crops�in particu­lar, improved drought tolerance and themore e�cient use of nitrogen. Ceres ener­gy crops are already on sale and several pi­lot projects that use them are under way.

Genomics, and the new biology it isbringing, thus promise a bright, practicalfuture. But some scientists wish to under­stand things merely for the joy of it. DavidHaussler, of the University of California,Santa Cruz, is one of them. Dr Hausslerwants to sequence 10,000 vertebrates, asixth of the total number of species of �sh,amphibians, reptiles, birds and mammals.Last month the BGI, in China, announcedit would take on the �rst 100 of these, fordelivery within two years.

Dr Haussler’s aim is to work out thecore vertebrate genome and see how it hasbeen modi�ed to produce the incrediblediversity of animals with backbones. TheGenome 10K project will, he reckons, cost$50m. That is not small change, but itamounts to only $5,000 a species, show­ing, once again, how the price of sequenc­ing has tumbled.

Dr Haussler has focused on vertebratesbecause he is one. Uncovering the geno­mic essentials of this ancient group wouldbe a coup. But genomics can also help toanswer more recent evolutionary ques­tions, in particular about how humansemerged and why they are unique. 7

THE decade since the genome an­nouncement has seen many remark­

able results. Vying with Dr Venter’s syn­thetic life for the title of the mostextraordinary was the announcement onFebruary 12th 2009 (by no mere coinci­dence Charles Darwin’s 200th birthday)that a second species of human had had itsgenome sequenced. Svante Paabo, the in­spiration for Michael Crichton’s novel and�lm, �Jurassic Park�, told a meeting of theAmerican Association for the Advance­ment of Science that his team at the MaxPlanck Institute in Leipzig had a version ofNeanderthal man’s DNA to compare withthat of modern humans.

The actual comparison was not pub­lished until six weeks ago, on May 6th. Itwas, however, worth waiting for. It

showed similarities between the species(in, for example, the FOXP2 gene that helpsgovern the ability to speak) as well as dif­ferences (in several genes connected withcognitive ability). These di�erences are ob­vious places to start looking for the essenceof modern humanity�the things that dis­tinguish Homo sapiens from other ani­mals, including other types of human, andthus accounts for the extraordinary �our­ishing of a species that is now estimated tomake use of 40% of the net primary pro­ductivity (the energy captured by photo­synthesis and converted into plant matter)of the planet’s land surface.

Genomics can, however, do more thanthis. Comparing the genomes of peoplealive today allows places where natural se­lection has been active in the more recent

past to be identi�ed. Until now this hasbeen done using gene chips, but the newpower of whole­genome sequencing willenrich and complete the picture. Alliedwith an analysis of how Homo sapiens hasspread out of Africa and around the world,this will give a �ne­grained picture of hu­man evolution�and possibly a controver­sial one if di�erences between geographi­cal groups emerge in sensitive areas suchas intelligence and behaviour. Such phe­notypic areas are also being investigated instudies looking at variation within popula­tions, for GWAS is being extended beyondthe realm of disease to examine intelli­gence, personality type, religiosity andeven the ability to make money.

If humanity has a historian­in­chief, hisname is Spencer Wells. Dr Wells is the head

The soul of an old machine

Genomics is raising a mirror to humanity, producing some surprising re�ections

12 A special report on the human genome The Economist June 19th 2010

2

1

of the Genographic Project, run jointly bythe National Geographical Society inWashington, DC, and IBM. In a sense, theproject is a type of personal genomics. Vol­unteers from all over the world (more than500,000 of them so far) give samples ofcheek cells for genetic analysis. In ex­change, they get a detailed breakdown ofthe wanderings of their ancestors over thepast 150,000 years or so.

Family treeFor Africans, or those of recent African ori­gin, the pattern is a tree leading out ofnorth­east Africa, where Homo sapiensseem to have originated, that spreads itsbranches through the whole continent. Forthose without recent African ancestry, thejourney then goes through a bottleneck60,000 years ago (a fact that was �rst hy­pothesised in the late 1980s, though it hastaken the Genographic Project’s plethoraof data to tie the date down exactly) beforefanning out into a tree that covers the restof the world.

The bottleneck corresponds to the pio­neers who crossed the straits of Bab elMandeb and populated the rest of theplanet (see map). And until recently therewas little interbreeding between thebranches, which has allowed local di�er­ences to emerge.

To start with, the spread of humanitywas followed by studying di�erences intwo unusual sorts of DNA: from cellularstructures called mitochondria whichhave their own genes and are inheritedonly from the mother, and from the part ofthe Y chromosome that passes only fromfather to son. Now, though, it is possible to

look at changes happening throughout thegenome.

Some of these changes will be random.Some, though, will be the product of natu­ral selection, and Pardis Sabeti of HarvardUniversity thinks she can work out whichis which. The blocks of DNA studied by theHapMap and similar projects are swappedbetween neighbouring chromosomes dur­ing the process of egg and sperm forma­tion. Dr Sabeti realised that a gene which isbeing favoured by selection will drag itsneighbours along for the ride, meaningthat the block it is in will be longer than thestatistics of random mixing would predict.

Using this and one or two similar statis­tical tools, Dr Sabeti has identi�ed 200places in the human genome that havebeen subject to recent selective �sweeps��and she often has a good idea of the genesthat have been doing the sweeping.

Some are not surprising. Genes that reg­ulate skin pigment and hair morphology,both well­known markers of geographicalorigin, have undergone signi�cant selec­tion. So have genes regulating metabolism,probably in response to the shift fromhunting and gathering to farming as man­kind’s principal way of life.

Intriguingly, though, several genes con­nected with sensory perception�hearingand balance, in particular�have altered. Inthis case, the changes are most noticeablein some of the Asian branches of human­ity. Genes involved in the development ofthe sound­detecting hair cells of the earsseem especially a�ected. Whether thatmeans Asians hear things di�erently fromother people has yet to be established, butit might.

The other sort of genes that haveevolved rapidly are those involved in in­fectious disease. Dr Sabeti has found evi­dence that malaria, tuberculosis, measlesand polio have all left their selective im­print as bits of DNA that help confer resis­tance are favoured. One of the biggest sur­prises was that Lassa fever, once regardedas a relatively rare disease, is also on thislist. Now, prompted partly by Dr Sabeti’swork, it is becoming established that Lassafever is actually very common in parts ofWest Africa. A �fth of Nigerians, for exam­ple, show signs of having been infected inthe past. That these people have not beenkilled by an illness originally thought to beup to 80% lethal suggests the protective ef­fect of the evolutionary change she has un­covered is strong.

Dr Sabeti has not, so far, found any se­lected genes that encode controversialtraits such as behaviour and intelligence.That may be because they have not beensubject to recent selection. It may, though,be because few links between these phe­notypes and the genes themselves haveyet been made.

That could change soon. As mentionedearlier, for example, the BGI in China plansa study to look speci�cally for intelligencegenes. And according to Nick Martin of theQueensland Institute of Medical Research,in Australia, dozens of GWAS investiga­tions of behavioural and cognitive linksare now under way. So as not to be caughtwith the dark­matter problem that has be­devilled the study of genetic disease, plansare afoot to consolidate these studies into�meta­analyses� big enough to detect oth­erwise­hidden e�ects. If all goes well, theold arguments about the role of nature andnurture in human development may soonbe seen in a new light.

That will also be true if Fred Gage, ofthe Salk Institute, in San Diego, is provedright in thinking that variation in cognitiveability may be genetically determined in away that is not inherited. Dr Gage studiesLINE­1elements, bits of DNA known collo­quially as jumping genes. These make upabout 20% of the human genome, andmost biologists do regard them as junk�orrather as parasites. Their origin is un­known, though they may be the distant de­scendants of some type of retrovirus (thesame class as the AIDS­causing virus, HIV)because they are able to copy themselvesusing a retroviral gene called reverse tran­scriptase. Generally, this is a bad thing fortheir hosts. If nothing else, it wastes re­sources. And if an element happens tocopy itself into the wrong place it can

The Economist June 19th 2010 A special report on the human genome 13

1

IT IS 2020. You are watching the latest epi­sode of CSI Miami. Horatio and the team

have a murder to solve. The murderer hasconveniently left a DNA sample behind. Infact, since a single strand of the moleculecan now be detected and analysed, hecould hardly avoid having done so. Not soconveniently, he is not on the database�wishy­washy civil libertarians having pro­hibited the collection of DNA recordsabout the unconvicted.

Never mind. Horatio pops the samplein a state­of­the­art sequencing machineand out comes a picture of what the sus­pect looks like�or, rather, a series of pic­tures of his likely appearance at �ve­yearintervals from age 15 to age 50. Cross­refer­ence these with Florida’s driving­licencedatabase, and the team has its man.

Not, perhaps, a nail­biting plot. But it is

a perfectly plausible description of the fu­ture of crime­�ghting. For in this and manyother ways, the development of genomicsmeans there will soon be no hiding place.

As Stewart Brand, an American futurol­ogist, memorably put it, �informationwants to be free,� and one of the lessons ofthe new biology is that it is all about infor­mation. DNA databases are a good illustra­tion of that, and of the con�icts and para­doxes the new age of genomics is creating.Everybody likes the idea of the guilty be­ing caught and punished. Universal DNA

databases would assist that process. Yetmany resist the logical conclusion thispoints to.

One reason they do so is an under­standable fear not so much of what gov­ernments might do with the informationnow as what they might do with it in the

future. DNA is more than just a reliable bio­metric (though not necessarily, if the priceof making it continues to fall, an unfake­able one). It is an individual’s essence. Ifanyone doubted that, Craig Venter’s ex­periment of implanting an arti�cial ge­nome in a cell he calls JCVI­syn1.0 and see­ing that cell’s daughters march to the newgenome’s tune should convince them.

Many people prefer to keep their es­sences to themselves. Few want theirweaknesses exposed to public gaze. Theymay not even want to confront thoseweaknesses in private (though this motiveis often less acknowledged). Yet that iswhat freedom of DNA information threat­ens. Disease susceptibility, life expectancy,personality traits, intelligence, criminaltendencies�all may be illuminated by theharsh light of free DNA information. Even

No hiding place

Everyday genomics is coming, ready or not

cause an existing gene to malfunction. Most biologists would have left things

at that, accepting that parasitism happensat all other levels of biology, so why not atthe level of the DNA? But Dr Gage won­dered if there was more to it. Five years agohe showed that LINE­1 elements are muchmore active in the brains of mice than intheir other tissues. And in 2009 he and acolleague, Nicole Coufal, showed the sameis true in people. Using sensitive sequenc­ing methods they discovered that humanbrain cells have around 100 more LINE­1el­ements in them than tissues such as theheart or the liver. That means the elementsare multiplying during the process of brainformation.

Survival of the �ttestWhy that is remains an unanswered ques­tion. But one possible explanation is this.Each new brain develops by a process akinto natural selection. Nerve cells grow anddevelop lots of connections with each oth­er, and then most of those connections andmany of the cells themselves die. Only the�ttest cells and connections survive, leav­ing a working network.

In this context a way of generating va­riations on which selection can operatemay actually result in better networks andthus brains. The random changes broughtabout by LINE­1 elements jumping around

the genomes of nerve cells can yield suchvariation. It is a daring hypothesis. Butthere is at least one other part of the body�the immune system�that relies on the in­ternal mutability of the DNA to generatethe variety needed for it to recognise all thepathogens that nature can throw at a body.Though the immune system has come upwith a di�erent solution in detail (particu­lar cassettes of DNA have evolved to shuf­�e round at this one place in the genome),the general principle is the same.

If Dr Gage’s hypothesis is correct, itleads to a nice irony. No one doubts that in­telligence has a heritable element. Studiesof identical and non­identical twins con­�rm that. No one doubts that upbringingand education matter, too. But part of thedi�erence between people’s cognitive abil­ities�between being a Darwin and adunce�might almost literally be a lottery,because it depends on the random move­ment of bits of DNA inside an individual’sdeveloping brain. 7

2

14 A special report on the human genome The Economist June 19th 2010

2

Previous special reports and a list offorthcoming ones can be found online

Economist.com/specialreports

Future special reportsDebt June 26thGambling July 10thEgypt July 17thLatin America September 11thForests September 25thThe world economy October 9th

Economist.com/rights

O�er to readersReprints of this special report are available at aprice of £3.50 plus postage and packing. A minimum order of �ve copies is required.

Corporate o�erCustomisation options on corporate orders of 100or more are available. Please contact us to discussyour requirements.

Send all orders to:

The Rights and Syndication Department26 Red Lion SquareLondon WC1R 4HQ

Tel +44 (0)20 7576 8148Fax +44 (0)20 7576 8492e­mail: rights@economist.com

For more information and to order special reportsand reprints online, please visit our website

if laws against genetic discrimination arepassed (as in America with the Genetic In­formation Nondiscrimination Act of2008), it is possible to imagine a future inwhich individuals are dogged by theirDNA. Would you turn it over to a putativespouse, for example?

Yet the bene�ts for medical research�and thus for the health of future genera­tions�of DNA information being free areenormous. And perhaps projects likeGeorge Church’s Personal Genome Projectwill show that those who are allowed tovolunteer their genes, rather than havingthem wrenched from them by the authori­ties, will be inclined to be generous. More­over, the unknown is often more terrifyingthan the known. Once the limits of DNA­based knowledge become apparent, someof the fears are likely to evaporate.

So much, then, for DNA freedom. Thosewho carelessly parrot Mr Brand, though,often forget that his quote actually startswith the words, �On the one hand infor­mation wants to be expensive because it isso valuable�. And the value of genomicscan be enormous.

The gods themselvesOn March 29th the Federal District Courtof New York ruled on a longstandingAmerican legal dispute. This was a claimby Myriad Genetics, of Salt Lake City, topatent protection on two human genescalled BRCA1 and BRCA2. Some versions ofthese genes increase the risk of breast can­cer, and Myriad sells a test that detectsthese versions. The patents in question,though, are not for the test but for the genesthemselves. Myriad claims to own the in­tellectual­property rights to these se­quences of DNA, even though they are nat­ural and found in every human being. Thecourt disagreed.

Dr Venter, meanwhile, is seeking a pat­ent on his newly minted DNA sequence forJCVI­syn1.0. He is on somewhat strongerground than Myriad, since the DNA inquestion is clearly an artefact, albeit onebased on a natural sequence. Also, bacteriaare not humans. Nevertheless, the princi­ple that anyone can �own� an organism’sDNA in this way disturbs many people.

What can and cannot be patentedneeds to be sorted out, for property rightslie at the heart of business. Patent law issupposed to encourage innovators, re­warding them with temporary monopo­lies but requiring them to place the detailsof their inventions in the public domainand thus open them to competitors. In thiscontext patenting an arti�cial genome for a

bacterium seems reasonable. As long asthe claims made are not too sweeping theyneed not stop anyone else patenting a dif­ferent arti�cial genome. (Patents on howsuch genomes are made might do so, butthat type of exclusivity is familiar territoryfor patent law.)

Similar rules should also apply to, say, acrop with an arti�cial genome. Such thingsboth need and deserve to be expensive.They need to be because they cost moneyto develop. They deserve to be because in­ventors, no less than authors, singers, ac­tors or any others whose work is easily andcheaply copied, are still worthy of theirhire. It should not apply, though, to a natu­ral genome.

Nor, many would argue, should it applyto synthetic DNA if that DNA is then insert­ed into a human being. Indeed, the ques­tion of whether such insertions should be

allowed at all is fraught. Once the recipe forhumanity is fully worked out, people willwant to try changing it. They will, nodoubt, plead medical need at the begin­ning but, as the rise of plastic surgery andthe modern debate about performance­enhancing drugs have shown, the medicalcan easily spill over into other areas.

Since ethical norms vary from countryto country, it is inconceivable that no onewill try this. If it works, it will almost cer­tainly spread. Thirty years ago the qualmsfelt by some about in­vitro fertilisationmelted in the face of the �rst gurgling childborn using the new technology. The sameis likely to happen for the �rst �enhanced�human, assuming the gurgling is happy.

As long as such changes remain withinthe human gene pool (wanting children tobe as tall, smart and beautiful as the best oftheir parents’ generation), a happy reac­tion is probably the right one. But what ifpeople want their children to be three me­tres tall and have blue skin, pointed earsand maybe even a tail? In practice, such�post­human� features would probably ar­rive gradually, giving people time to adjustto the idea and decide which, if any, wereacceptable. Some might be accepted. Mostwill probably be rejected as monstrous.For, in the end, people usually manage tobend technology to their will, rather thanthe other way round.

There will be mistakes on the way, andsu�ering, too. But technology, once invent­ed, cannot be unlearned. Perhaps, then,the last word belongs to Mr Brand. Whenhe set up the Whole Earth Catalog, the ven­ture that �rst brought him to public atten­tion, he said �We are as gods, and might aswell get good at it.� 7