The Seed State of Science 2008

The inaugural edition of Seed's annual State of Science explores the current scientific landscape, from publishing to public perception, and profiles emergent hotspots.

Read more Seed State of Science 2008


Bigger Faster Better

By Craig Venter | Posted November 20, 2008

MetricGrading Science

Grading Science

PISA is a standardized test administered by the OECD that attempts to measure the abilities of 15-year-olds in three categories: identifying scientific issues, explaining phenomena scientifically, and using scientific evidence. Despite being one of the biggest spenders on education, in terms of both total dollars and percentage of GDP, the United States ranked 35 out of the 57 countries where the test was given. Metric source: PISA 2006 Executive Summary, OECD Factbook 2008.

Big science has not always proven to be good science. As the big push into genomics began, backed by billions of dollars in government money, a slew of new genome centers opened in the US and Japan, and the Wellcome Trust in the UK funded the Sanger Center. Government funding is often driven by social policy in place of science. The yeast genome effort, for example, involved 10 years and approximately 1,000 scientists spread around the European Union and the US. The goal of much of the funding was to expand the use of the sequencing and informatics technology across the EU, but this meant that quality couldn't be controlled; the first published yeast chromosome had to be re-sequenced. Nevertheless this was the model that drove the international human genome project.

It is no wonder that it was assumed that armies of scientists and technicians would be required to sequence the human genome. The thinking was based on how much DNA sequence could be produced per day by individual investigators using what are now considered primitive tools. But we smashed that old paradigm in 1995 when my relatively small team sequenced the first genome of a living species, that of Haemophilus influenzae, in just four months. We used a new method we named whole-genome shotgun sequencing, which worked by randomly selecting and sequencing DNA fragments from the organism's genome, then computationally reassembling it.

Our success came despite a lack of NIH funding after it insisted that the method was "unlikely to work." We relied on our funding from Human Genome Sciences to do the experiment, and it was then only possible to carry out the work because the money HGS had given to TIGR had no strings attached—something that they regretted at the time because they, too, did not want the experiment done. Such independence has continued to be an essential aspect of my science's progress.

After the publication of that first genome, government money to sequence more and more microbial genomes poured into my institute. I established a technology core to sequence, assemble, and annotate genomes in an assembly-line manner. This ensured consistently high-quality data that was independent of any individual scientist's skill set. The method was far superior to other projects. Any scientist or any collaborator could come to TIGR, have a microbial genome sequenced and analyzed, and produce the highest-quality publications. The constant challenge for me was to push the scientists to do more than just data collection. We had made it easy to sequence genomes, but I did not want TIGR to be just a sequencing factory.

From that first genome, H. influenzae, I drove the team to try to understand basic life of a species by viewing its entire genome. Sometimes just the selection of which species genome to sequence guaranteed exciting outcomes, such as when we sequenced the smallest known genome, Mycoplasma genitalium, and the first archaea, Methanococcus jannaschii, which was done in collaboration with Carl Woese to help prove that his third branch of life was real. Comparison of the M. genitalium and H. influenzae genomes raised an interesting question: If one species needed 1,800 genes and the other only 500 or so, was there a minimal genome that could sustain independent self-replicating life? Years of comparative genomics and gene knockouts have not answered completely the question, so we have turned to synthesizing genomes and varying their contents so we can find out just how little genetic information is necessary to make an organism work.

Despite my urging that we always look for those big questions, data generation for its own sake continues to be a major impediment to real scientific breakthroughs in genomics. It is not hard to understand why investigators, particularly young scientists, are satisfied being data generators, as government agencies and some foundations continue to pay out hundreds of millions of dollars for just DNA sequencing or, even worse, microarrays, creating huge datasets but seldom any real scientific insight.

I was certain, after all of our success with our new whole-genome shotgun sequencing method, that the Human Genome Project would adopt our methods, but the federal human genome project was bureaucratically moribund and unable to change. In truth there was no incentive to change. The scientists in the various genome centers benefited hugely— some on the order of $100 million per year or more— from a process that was essentially a decentralized, widely distributed public-works project.

When it became clear that my approach would not be adopted, I resigned myself to the fact that I would not be working on the human genome. Applied Biosystems (ABI), an instrument company, then came to me with a new automated machine for DNA sequencing that they thought would work well with my shotgun method. They were interested in funding an independent effort to sequence the human genome and using it as an opportunity to show off their product, much like how car companies sponsor racing teams, or IBM creates computers to play chess or do big calculations for publicity. The only difference was that ABI wanted to start a new company around sequencing the genome, not just to sponsor the work.

I accepted the challenge and Celera Genomics was formed. We had initially hoped for cooperation and even collaboration with the public effort, but we were only seen as a threat to the status quo of a mega-funded, mega-distributed project, and the effort quickly turned to a competition. Thankfully though, competition can be healthy for science and heighten its public impact. Competition keeps scientists and the science honest—you want to be first, and you need to be right. Competition was clearly good for the Human Genome Project. By even the most critical accounts of the effort, the initial goal—having the first draft of the human genome sequence— happened years earlier than had been expected, changing the face of research seemingly overnight.


Bigger Faster Better
By Craig Venter
Posted November 20, 2008
Originally appeared in Seed 19

SEEDMAGAZINE.COM by Seed Media Group. ©2005-2009 Seed Media Group LLC. All Rights Reserved.

Sites by Seed Media Group: Seed Media Group | ScienceBlogs | SEEDMAGAZINE.COM