TIM O'REILLY
Founder, O'Reilly Media, O'Reilly Radar

The Quantitative Revolution in Biology

George Church asked "Is life a qualitative or quantitative question?" Every revolution in science has come when we learn to measure and count rather than asking binary qualitative questions. Church didn't mention phlogiston, but it's what came to mind as a good analogy. Heat is not the presence or absence of some substance or quality, but rather a measurable characteristic of a complex thermodynamic system. Might not the same be true of life?

The measurement of self-replication as a continuum opens quantitative vistas. Here are a few tidbits from George Church and Craig Venter:

• The most minimal self-replicating system measured so far has 151 genes; bacteria and yeast about 4000; humans about 20,000.

• There are 12 possible amino acid bases (6 pairs); we ended up using 4 bases (2 pairs); other biological systems are possible.

• Humans are actually an ecology, not just an organism. The human microbiome: 23K human genes, 10K bacterial genes.

• Early estimates of the number of living organisms were limited to those that could be cultured in the laboratory; by sampling the DNA in water and soil, we have discovered that we undercounted by many orders of magnitude

• The biomass of bacteria deep in the earth is greater than the biomass of all visible plants and animals; ditto the biomass of ocean bacteria.

• The declining cost of gene sequencing is outpacing Moore's Law (1.5x/year): the number of base pairs sequenced per dollar is increasing at 10x per year.

Net: The current revolution in genomics and synthetic biology will be as profound as the emergence of modern chemistry and physics from medieval alchemy.

ED REGIS
Writer; What Is Life?

Almost fifteen years ago, in a profile of Leroy Hood, I quoted Bill Gates, who said: "The gene is by far the most sophisticated program around."

At the Edge Master Class last weekend I learned the extent to which we are now able to reprogram, rework, and essentially reinvent the gene. This gives us a degree of control over biological organisms — as well as synthetic ones — that was considered semi-science fictional in 1995. Back then scientists had genetically engineered E. coli bacteria to produce insulin. At the Edge event, by contrast, Craig Venter was talking about bacteria that could convert coal into methane gas and others that could produce jet fuel. It was merely a matter of doing the appropriate genomic engineering: by replacing the genome of one organism with that of another you could transform the old organism into something new and better.

George Church, for his part, described the prospect of synthetic organisms grown from mirror-image DNA; humanized mice, injected with human genes so that they would produce antibodies that the human body would not reject; and the possibility of resurrecting extinct species including the woolly mammoth and Neanderthal man.

But as far-out as these developments were, none of them was really surprising. After all, science and technology operate by systematically gaining knowledge of the world and then applying it intelligently. Thus we skip from miracle to miracle.

More extraordinary to me personally was the fact that the first day of the EDGE event was being held on the premises of a private rocket manufacturing facility in Los Angeles, SpaceX, which also builds Tesla electric vehicles, all under the leadership of Internet entrepreneur Elon Musk. The place was mildly unbelievable, even after having seen it with my own eyes. In the age of Big Science, where it is not uncommon for scientific papers to be written by forty or more coauthors, the reign of the individual is not yet dead.

VICTORIA STODDEN
Computational Legal Scholar, Yale Law School

Craig Venter posed the question whether it is possible to reconstruct life from its constituent parts. Although he's come close, he hasn't done it (yet?) and neither has anyone else. Aside from the intrinsic interest of the question, its pursuit seems to be changing biological research in two fundamental ways encapsulated Venter's own words:

We have these 20 millions genes. I view these as design components. We actually have software now for designing species, where we can try and put these components together. The biggest problem with engineering biology on first principles is that we don't know too many first principles. It's a minor problem! In fact, from doing this, if we build this robot that can make a million chromosomes a day, and a million transformations, and a million new species versions, it'll be the fastest way to establish what the first principles are, if we can track all that information and all the changes.           

Unlike physics or more mathematical fields, research in biology traditionally hasn't been a search for underlying principles, or had the explicit goal of developing grand unifying theories. A cynic could even argue funding incentives in biology encourage complexity: big labs are funded if they address very complicated, and thus more expensive to research, phenomena. Whether or not that's true, chemical reconstruction of the genome is a process from first principles, marking a change in approach that brings biological research closer in spirit to more technical fields. Venter seems to believe that answering questions such as, "Can we reconstruct life from its components?" "What genes are necessary for life?" "What do you really need to run cellular machinery?" and "What is a minimal organism that could survive?" will uncover first principles in biology, potentially structuring understanding deductively.            
Venter's use of combinatorial biological research is another potential sea-change in the way understanding is developed. This use of massive computing is analogous to that occurring in many other areas of scientific research, and the key is that discovery becomes less constrained by a priori assumptions or models (or understanding?).

Moore's Law and ever cheaper digital storage is giving scientists the luxury of solution search within increasingly large problem spaces. With complete search over the space of all possible solutions, in principle it is no longer necessary to reason one's way to the (an?) answer. This approach favors empirical evaluation over deductive reasoning. In Venter's biological context, presumably if automated search can find viable new species it will then be possible to investigate their unique life enabling characteristics. Perhaps through automated search?

JESSE DYLAN
Film-Maker, Form.tv, FreeForm.tv

What a revelation the The Master Class in Synthetic Genomics was. In addition to being informative on so many literal levels it reinforced the mystery and wonder of the world. George Church and Craig Venter were generous to give us a glimpse of where we are today and fire the imagination of where we are going. It's all science but seems beyond science fiction — living forever, reprogramming genes, resurrecting extinct species. All told at SpaceX — a place where people are reaching for the stars, not just thinking about it but building rockets to take us there. Where Elon Musk contemplates the vastness of space and our tiny place in it, where we gained a perspective on the things that are very small and beyond the vision of our eyes. So small it's a wonder we even know they are there. Thanks for giving us a profound glimpse into the future.

We are in such an early formative stage it makes one wonder where we will be in a hundred or even a thousand years. It's nice to be up against mysteries.

GEORGE DYSON
Science Historian; Darwin Among the Machines

End Of Species

We speak of reading and writing genomes — but no human mind can comprehend these lengthy texts. We are limited to snippet view in the library of life.

As Edge's own John Markoff reported from the recent Asilomar conference on artificial intelligence, the experts "generally discounted the possibility of highly centralized superintelligences and the idea that intelligence might spring spontaneously from the Internet."

Who will ever write the code that ignites the spark? Craig Venter might be hinting at the answer when he tells us that "DNA... is absolutely the software of life." The language used by DNA is much closer to machine language than any language used by human brains. It should be no surprise that the recent explosion of coded communication between our genomes and our computers largely leaves us out.

"The notion that no intelligence is involved in biological evolution may prove to be one of the most spectacular examples of the kind of misunderstandings which may arise before two alien forms of intelligence become aware of one another," wrote viral geneticist (and synthetic biologist) Nils Barricelli in 1963. The entire evolutionary process "is a powerful intelligence mechanism (or genetic brain) that, in many ways, can be comparable or superior to the human brain as far as the ability of solving problems is concerned," he added in 1987, in the final paper he published before he died. "Whether there are ways to communicate with genetic brains of different symbioorganisms, for example by using their own genetic language, is a question only the future can answer."

We are getting close.

ALEXANDRA ZUKERMAN
Assistant Editor, Edge

As the meaning of George Church and Craig Venter's words permeated my ever-forming pre-frontal cortex at The Master Class, I cannot deny that I felt similarly to the way George Eliot described her own emotions in 1879. Eliot, speaking as Theophrastus in a little-known collection of essays published that year, predicts that evermore perfecting machines will imminently supercede the human race in "Shadows of the Coming Race:"

When, in the Bank of England, I see a wondrously delicate machine for testing sovereigns, a shrewd implacable little steel Rhadamanthus that, once the coins are delivered up to it, lifts and balances each in turn for the fraction of an instant, finds it wanting or sufficient, and dismisses it to right or left with rigorous justice; when I am told of micrometers and thermopiles and tasimeters which deal physically with the invisible, the impalpable, and the unimaginable; of cunning wires and wheels and pointing needles which will register your and my quickness so as to exclude flattering opinion; of a machine for drawing the right conclusion, which will doubtless by-and-by be improved into an automaton for finding true premises — my mind seeming too small for these things, I get a little out of it, like an unfortunate savage too suddenly brought face to face with civilisation, and I exclaim —

'Am I already in the shadow of the Coming Race? and will the creatures who are to transcend and finally supersede us be steely organisms, giving out the effluvia of the laboratory, and performing with infallible exactness more than everything that we have performed with a slovenly approximativeness and self-defeating inaccuracy?' 1

Whereas Theophrastus' friend, Trost (a play on Trust) is confident that the human being is and will remain the "nervous center to the utmost development of mechanical processes" and that "the subtly refined powers of machines will react in producing more subtly refined thinking processes which will occupy the minds set free from grosser labour," Theophrastus feels "average" and less energetic, readily imagining his subjugation by these steely organisms giving out the "effluvia of the laboratory." He imagines instead that machines operate upon him, measuring his thoughts and quickness of mind. Micrometers, thermopiles and tasimeters were invading the sanctity of his consciousness with their "unconscious perfection."

As George Church told us that "We're getting to a point where we can really program these cells as if they were an extension of a computer" and "This software builds its own hardware — it turns out biology does this really well," my sensibilities felt slightly jarred. Indeed, I felt as though I might be from an uncivilized time and place, suddenly finding myself on the platform as a flying train whizzed past (in fact, our tour of SpaceX and Tesla by Elon Musk was not far off!).

I asked myself the same question as George Eliot posed to herself over one hundred years ago: If computing and genetics are converging, such that computers will be reading our genomes and perfecting them, has not Eliot's prediction come true? I wondered, as a historian of science, not as much about the implications of such a development, but more about why computers have become so powerful. Why do we trust, as Trost does, artificial intelligence so much? Will scientists ultimately give their agency over to computers as we get closer to mediating our genomes and that of other forms of life? Will computers and artificial intelligence become a new "invisible hand" such as that which guides the free market without human intervention? I am curious about the role computers will be playing, as humans grant them more and more hegemony.

___

1 George Eliot, Impressions of Theophrastus Such

LAWRENCE KRAUSS
Physicist, Director, Origins Initiative, ASU; Author, Hiding In The Mirror

What struck me was the incredible power that is developing in bioinformatics and genomics, which so resembles the evolution in computer software and hardware over the past 30 years.

George Church's discussion of the acceleration of the Moore's law doubling time for genetic sequencing rates,, for example, was extraordinary, from 1.5 efoldings to close to 10 efoldings per year. When both George and Craig independently described their versions of the structure of the minimal genome appropriate for biological functioning and reproduction, I came away with the certainty that artificial lifeforms will be created within the next few years, and that they offered great hope for biologically induced solutions to physical problems, like potentially buildup of greenhouse gases.

At the same time, I came away feeling that the biological threats that come with this emerging knowledge and power are far greater than I had previously imagined, and this issue should be seriously addressed, to the extent it is possible. But ultimately I also came away with a more sober realization of the incredible complexity of the systems being manipulated, and how far we are from actually developing any sort of comprehensive understanding of the fundamental molecular basis of complex life. The simple animation demonstrated at the molecular level for Gene expression and replication demonstrated that the knowledge necessary to fully understand and reproduce biochemical activity in cells is daunting.

Two other comments: (1) was intrigued by the fact that the human genome has not been fully sequenced, in spite of the hype, and (2) was amazed at the available phase space for new discovery, especially in forms of microbial life on this planet, as demonstrated by Craig in his voyage around the world, skimming the surface, literally, of the ocean, and of course elsewhere in the universe, as alluded to by George.

Finally, I also began to think that structures on larger than molecular levels may be the key ones to understand for such things as memory, which make the possibilities for copying biological systems seem less like science fiction to me. George Church and I had an interesting discussion about this which piqued my interest, and I intend to follow this up.

Return to "A Short Course On Synthetic Genomics"