Lab life 3.0: synthetic biology's long experiment in governing itself

Synthetic biology has always numbered its big meetings like software releases. The first international conference, held at MIT in 2004, was Synthetic Biology 1.0. Berkeley hosted 2.0 in May 2006, and in June 2007 several hundred researchers, investors and company founders met at ETH Zurich for 3.0. The numbering carried a quiet promise that each version would fix the bugs of the last.

Read back through the record of those early years and one bug stands out, because it never really got patched. It wasn't technical. It was the question of who writes the rules for a field whose whole ambition is to make living things easier to engineer, and whether the engineers should be the ones holding the pen.

Seen from May 2024, that question looks different than it did in Zurich. Not as different as you might expect.

A rulebook drafted in the room

By 2006 synthetic biologists had reason to worry about their public image. In 2002 a team at Stony Brook University had assembled infectious poliovirus from synthetic DNA. In 2005 researchers reconstructed the 1918 pandemic influenza virus. Commercial gene synthesis was getting cheaper every year, and journalists had begun asking the obvious question: could someone simply order a dangerous genome by mail?

Ahead of SB2.0, organizers circulated a draft declaration on community self-governance. Its proposals were sensible enough. Gene synthesis companies would check orders against lists of hazardous sequences, researchers would build better software to flag them, and the community would talk openly about risk. The draft was also, fairly obviously, a message to governments that outside rules were unnecessary.

The response from outside the field came quickly. Thirty-eight civil society organizations signed an open letter rejecting the idea that a young industry should define its own oversight. Their objection was less to screening than to the frame around it. Biosecurity, they argued, was one question among many. Others included patents on basic biological parts, the effect on farmers if factories learned to make what their crops supplied, the release of novel organisms into the environment, and the plain fact that the people most affected had no seat in the room. A rulebook written by practitioners, at a meeting its critics could not get into, looked less like governance than like an attempt to head it off.

The declaration was never adopted as a formal community position. The instinct behind it survived intact.

Asilomar, and the half of the story people remember

Whenever scientists propose to govern themselves, someone mentions Asilomar. In February 1975 around 140 people, most of them molecular biologists, gathered at the Asilomar conference grounds on the California coast to decide what to do about recombinant DNA. A year earlier a group led by Paul Berg had asked colleagues to pause certain experiments voluntarily. At Asilomar the participants agreed to match experiments with levels of physical and biological containment according to how risky they seemed. In 1976 the US National Institutes of Health turned that approach into formal guidelines, overseen by its Recombinant DNA Advisory Committee.

The meeting deserves much of its reputation. A research community spotted a hazard, slowed down and built a system before anything went badly wrong. Plenty of later technologies never got that much forethought.

Asilomar also had limits, several of which its own organizers later acknowledged. The agenda was confined to laboratory safety; the ethics of genetic engineering and its possible military uses were set aside. The people deciding were nearly all scientists, with a handful of lawyers and journalists. And the NIH guidelines applied to federally funded research, which left private companies formally outside them at exactly the moment biotechnology was turning into an industry.

When synthetic biologists cite Asilomar, they usually mean the first part: responsible scientists acting early. When critics cite it, they mean the second: a narrow agenda, set by insiders. Both readings are accurate. The mistake is treating either one as the whole story.

The part that stuck: checking DNA orders

The idea from 2006 that did take root was screening. In 2007, around the time of the Zurich meeting, a commentary in Nature Biotechnology set out a framework from a newly formed trade body, the International Consortium for Polynucleotide Synthesis, written together with academic researchers and US security officials. Providers would screen orders for sequences of concern and share the tools to do it; customers would identify themselves and their institutions. Critics said the framework dealt with deliberate misuse and nothing else. That was true, and its authors would probably not have argued.

What followed was a slow ratchet:

• 2009. Several of the largest commercial gene synthesis firms formed the International Gene Synthesis Consortium and published a Harmonized Screening Protocol: check every order against databases of regulated pathogens and toxins, vet customers, keep records, report suspicious requests. A separate industry association published its own code of conduct the same year.
• 2010. The US Department of Health and Human Services issued voluntary screening guidance for providers of synthetic double-stranded DNA, broadly in line with what the larger firms were already doing.
• 2017. The consortium revised its protocol.
• October 2023. HHS replaced its 2010 guidance with a broader version covering all synthetic nucleic acids, including benchtop synthesis machines, and asking providers to screen progressively shorter sequences over the following three years. The same month, President Biden's executive order on artificial intelligence directed the White House science office to produce a screening framework, tying the issue to fears that AI tools could make biological design easier.
• April 2024. That framework arrived. Its lever is money: researchers holding federal grants are expected to buy synthetic DNA only from providers that attest to following screening practices, with the requirements phased in over the coming years.

So in under two decades a voluntary industry habit has become something close to a requirement, enforced through federal purchasing rather than law. That is a genuine achievement, and it happened partly because the industry did the groundwork.

A fair ledger

It is easy to be cynical about self-governance, and easy to be credulous. A ledger helps.

What it delivered:

1. A working norm. Among the major providers, screening is routine. Someone ordering from a consortium member in 2024 should expect both the sequence and the customer to be checked.
2. Shared tools. Common protocols and databases, and more recently free screening software from non-profit groups, which lets smaller companies take part without building their own systems.
3. A draft for government. When Washington finally moved, it didn't have to invent screening from scratch. It wrote down practice that already existed and attached conditions.
4. Working relationships. Industry and security agencies talk regularly. That is easy to dismiss and hard to replace.

What it did not:

1. Universal coverage. Membership is voluntary and there is no global register of DNA providers. Firms outside the consortium, especially in countries without rules of their own, set their own practices. Benchtop synthesizers, which move DNA production into individual labs, strain a model built around a few central suppliers.
2. Accountability. Members attest that they follow the protocol. There is no public audit and no penalty beyond reputation if screening fails. The April 2024 framework adds a market incentive, but not an inspector.
3. Scope. This is the big one. Almost everything in the 2006 open letter that wasn't biosecurity found no home in self-governance. Ownership and patents, demand for biomass, environmental release, the livelihoods of commodity farmers: these were handled, if at all, somewhere else. The most persistent venue has been the UN Convention on Biological Diversity, where synthetic biology has been on the agenda since 2010 and where, in December 2022, governments agreed to set up a regular process for horizon scanning and assessment. It is slow, consensus-bound diplomacy, but it is a place where farmers' organizations and governments from the global South can speak.
4. A wider table. The conferences did open up a little. By SB4.0 in Hong Kong in 2008, civil society groups had a panel on the program, and the iGEM student competition later built "human practices" work into how teams are judged. But the rulemaking itself stayed with practitioners, funders and security agencies.

What the Zurich argument looks like now

The 2007 critique of self-governance has aged unevenly. Its central prediction, that the industry would manage the risks it found convenient and leave the rest alone, has largely held. Its suggestion that screening was mainly public relations has not; screening is real, imperfect and useful, and governments are now building on it rather than replacing it.

The framing has also shifted. In 2007 the argument was self-governance versus regulation, as if one had to win. In 2024 the more accurate picture is a set of layers: industry practice at the bottom, government conditions attached to funding and procurement above it, and slow international discussion over the top. Each layer covers different ground. The weakness sits in the questions that none of them covers: the economic and social ones a 2006 declaration left out, and that nobody since has had a clear mandate to answer.

Europe has its own version of this gap. The debate over deregulating plants made with new genomic techniques has turned, again and again, on whether a safety assessment alone counts as governing a technology. And the ownership questions raised in Zurich, including patent claims on what Craig Venter's team hoped would be the first synthetic organism, never went away. They simply moved from conference halls to patent offices.

Asilomar remains a good story about scientists taking responsibility for laboratory hazards. It is a weaker template for deciding what a technology is for and who benefits, because nobody at Asilomar was asked that. Eighteen years after Berkeley, synthetic biology still has no forum where that question is the main item on the agenda.