Researchers produce first complete computer model of an organism

July 20, 2012 By Max McClure, Stanford University

The Covert Lab incorporated more than 1,900 experimentally observed parameters into their model of the tiny parasite Mycoplasma genitalium. Illustration: Erik Jacobsen / Covert Lab
( -- In a breakthrough effort for computational biology, the world's first complete computer model of an organism has been completed, Stanford researchers reported in the journal Cell.

A team led by Stanford Professor Markus Covert used data from more than 900 scientific papers to account for every that takes place in the of Mycoplasma genitalium – the world's smallest free-living bacterium.

By encompassing the entirety of an organism in silicon, the paper fulfills a longstanding goal for the field. Not only does the model allow researchers to address questions that aren't practical to examine otherwise, it represents a stepping-stone toward the use of computer-aided design in bioengineering and medicine.

"This achievement demonstrates a transforming approach to answering questions about fundamental biological processes," said James M. Anderson, director of the National Institutes of Health Division of Program Coordination, Planning and Strategic Initiatives. "Comprehensive computer models of entire cells have the potential to advance our understanding of cellular function and, ultimately, to inform new approaches for the diagnosis and treatment of disease."

The research was partially funded by an NIH Director's Pioneer Award from the National Institutes of Health Common Fund.

From information to understanding

Biology over the past two decades has been marked by the rise of high-throughput studies producing enormous troves of cellular information. A lack of experimental data is no longer the primary limiting factor for researchers. Instead, it's how to make sense of what they already know.

Most biological experiments, however, still take a reductionist approach to this vast array of data: knocking out a single gene and seeing what happens.

"Many of the issues we're interested in aren't single-gene problems," said Covert. "They're the complex result of hundreds or thousands of genes interacting."

This situation has resulted in a yawning gap between information and understanding that can only be addressed by "bringing all of that data into one place and seeing how it fits together," according to Stanford bioengineering graduate student and co-first author Jayodita Sanghvi.

Integrative computational models clarify data sets whose sheer size would otherwise place them outside human ken.

"You don't really understand how something works until you can reproduce it yourself," Sanghvi said.

Small is beautiful

Mycoplasma genitalium is a humble parasitic bacterium, known mainly for showing up uninvited in human urogenital and respiratory tracts. But the pathogen also has the distinction of containing the smallest genome of any free-living organism – only 525 genes, as opposed to the 4,288 of E. coli, a more traditional laboratory bacterium.

Despite the difficulty of working with this sexually transmitted parasite, the minimalism of its genome has made it the focus of several recent bioengineering efforts. Notably, these include the J. Craig Venter Institute's 2009 synthesis of the first artificial chromosome.

"The goal hasn't only been to understand M. genitalium better," said co-first author and Stanford biophysics graduate student Jonathan Karr. "It's to understand biology generally."

Even at this small scale, the quantity of data that the Stanford researchers incorporated into the virtual cell's code was enormous. The final model made use of more than 1,900 experimentally determined parameters.

To integrate these disparate data points into a unified machine, the researchers modeled individual biological processes as 28 separate "modules," each governed by its own algorithm. These modules then communicated to each other after every time step, making for a unified whole that closely matched M. genitalium's real-world behavior.

Probing the silicon cell

The purely computational cell opens up procedures that would be difficult to perform in an actual organism, as well as opportunities to reexamine experimental data.

In the paper, the model is used to demonstrate a number of these approaches, including detailed investigations of DNA-binding protein dynamics and the identification of new gene functions.

The program also allowed the researchers to address aspects of cell behavior that emerge from vast numbers of interacting factors.

The researchers had noticed, for instance, that the length of individual stages in the cell cycle varied from cell to cell, while the length of the overall cycle was much more consistent. Consulting the model, the researchers hypothesized that the overall cell cycle's lack of variation was the result of a built-in negative feedback mechanism.

Cells that took longer to begin DNA replication had time to amass a large pool of free nucleotides. The actual replication step, which uses these nucleotides to form new DNA strands, then passed relatively quickly. Cells that went through the initial step quicker, on the other hand, had no nucleotide surplus. Replication ended up slowing to the rate of nucleotide production.

These kinds of findings remain hypotheses until they're confirmed by real-world experiments, but they promise to accelerate the process of scientific inquiry.

"If you use a model to guide your experiments, you're going to discover things faster. We've shown that time and time again," said Covert.


Much of the model's future promise lies in more applied fields.

CAD – computer-aided design – has revolutionized fields from aeronautics to civil engineering by drastically reducing the trial-and-error involved in design. But our incomplete understanding of even the simplest biological systems has meant that CAD hasn't yet found a place in bioengineering.

Computational models like that of M. genitalium could bring rational design to biology – allowing not only for computer-guided experimental regimes, but for the wholesale creation of new microorganisms.

Once similar models have been devised for more experimentally tractable , Karr envisions bacteria or yeast specifically designed to mass-produce pharmaceuticals.

Bio-CAD could also lead to enticing medical advances – especially in the field of personalized medicine. But these applications are a long way off, the researchers said.

"This is potentially the new Human Genome Project," Karr said. "It's going to take a really large community effort to get close to a human model."

Explore further: 'Virtual cell' could bring benefits of simulation to biology

More information:

Related Stories

Biologists use computers to study bacterial cell division

January 25, 2008

A group of computational biologists at Virginia Tech have created a mathematical model of the process that regulates cell division in a common bacterium, confirming hypotheses, providing new insights, identifying gaps in ...

Researchers make the leap to whole-cell simulations

March 30, 2011

Researchers have built a computer model of the crowded interior of a bacterial cell that – in a test of its response to sugar in its environment – accurately simulates the behavior of living cells.

Recommended for you

Paleontologists report world's biggest Tyrannosaurus rex

March 22, 2019

University of Alberta paleontologists have just reported the world's biggest Tyrannosaurus rex and the largest dinosaur skeleton ever found in Canada. The 13-metre-long T. rex, nicknamed "Scotty," lived in prehistoric Saskatchewan ...

NASA instruments image fireball over Bering Sea

March 22, 2019

On Dec. 18, 2018, a large "fireball—the term used for exceptionally bright meteors that are visible over a wide area—exploded about 16 miles (26 kilometers) above the Bering Sea. The explosion unleashed an estimated 173 ...


Adjust slider to filter visible comments by rank

Display comments: newest first

3 / 5 (2) Jul 20, 2012
Great news! But I'm a bit skeptical of the writer claim: the model will "account for every molecular interaction that takes place in the life cycle of Mycoplasma genitalium." Let's review the paper.
4.2 / 5 (5) Jul 20, 2012
i understand how a car works but i wouldn't be able to build one

Grab a box of meccanos or lego technics, and I'm pretty sure you could make something that has the basic functions and mechanisms of a car - engine, gearbox, suspension, steering, four wheels at the corners. You could run it on pressurized air.

That would be a similiar model of a car, as the silicon chip is a model of the bacterium.
5 / 5 (3) Jul 20, 2012
@ Bowler

You understand what a car does -- not how it works as you stated.

there are many users of computers -- but it takes a electrical engineer to understand how it works -- a CS major to understand how its programming works -- a physicist to understand how its individual components work.

You merely understand what a car does -- it moves -- you don't understand the mathematics of it's movement until you can start from scratch and mimic it perfectly.

-- to build a computer simulation of the complete lifecycle of an organism and it to be accurate down to every chemical interaction means they have complete understanding the chemistry and biology of the organism.

They are modeling to recreate all the phenotypes (traits) of the organism based on idividual molecules and their interactions.

* Please note this has nothing to do with tracking every organelle or even DNA/RNA coping, fixing, or translation. But its part of those functions.
5 / 5 (1) Jul 20, 2012
the test would be to have a colony of virtual Mycoplasma and provide it with the same conditions as a real colony. the test would be how similar they are after a few hundred generations.
5 / 5 (1) Jul 20, 2012
I have reviewed the paper and tried to see if the model "account for every molecular interaction that takes place in the life cycle of Mycoplasma genitalium."

Well, it sounds like an awesome model and I agree it will be breakthrough tool in the discovery of the biology of cells and organisms in general.

But, not knowing the field at all, I expected to hear about interaction between every molecule. It is not. It is a model of major interactions or of major processes at molecular scale. They do not model every molecule in the cell; they model all major processes in the cell.

But nevertheless they create a model that can generate new emergent properties of cell function based on the mass of basic processes in a cell.

I would not say that they reached the holy grail of the biology. That's still way off where each molecule would be modeled separately, but really great milestone/breakthrough.
5 / 5 (2) Jul 20, 2012
... and Go Stanford!
5 / 5 (2) Jul 20, 2012
... and Go Stanford!

Amen, my Friend! Stanford ROCKS and you know that...

5 / 5 (2) Jul 20, 2012
This is GREAT news and very exciting! ...and the organism studied is only a factor of 8 smaller that it's big brothers in the bacterial world. It is just a short wait until all the bacterial sized critters can be modeled and then we're off moving towards modeling the big multicellular organisms. GO ARTIFICIAL LIFE! EVOLVE before our very eyes!!!
1 / 5 (1) Jul 20, 2012
there are many users of computers -- but it takes a electrical engineer to understand how it works -- a CS major to understand how its programming works -- a physicist to understand how its individual components work.

Actually, you don't need all those titles to make a computer. You just have to have a good grasp of college level math in first order logic, and understand the concept of a state machine. Of course you need to know how to hold a soldering iron to actually build it, but a basic computer can be made with about 50-100 electromechanical relays that are simple enough that you can understand and make them yourself without a degree in physics. The basic components are an adder, a memory register or two, and some sort of data bus to switch inputs and outputs. The rest is math and logic.

And programming the thing is just about what you want it to do. Making the relay computer count to ten would be simple, since you designed the instruction set to start with.
not rated yet Jul 20, 2012
Beautiful, but it's maybe step 3,012 of a few million. Imagine how hard quorum sensing would be, needing perhaps dozens or hundreds of individual cells.
3 / 5 (2) Jul 20, 2012
Actually, this is Noble prize grade stuff.

Not that it would be unfathomably hard to do, but because this is the right thing to do right now. Having established this feat is a milestone, after which the next groups can confidently attack simulating larger organisms.

This represents a turning point in simulating living things. From this day on, it is just a matter of increasing computational capacity, before we can simulate a vertebrate, or ultimately, an entire human being. (Mind you, this is only on the cellular level. Nothing to do with psychology or such stuff, for some time yet.)

This is the kind of science that we have just begun to do, and these are the first guys to do it. Congratulations!
5 / 5 (1) Jul 20, 2012
It is also great for astrobiology! Of course the major processes will differ, but the regulatory interplay is a lot easier to get to than trying to model bottom up chemical networks. (Especially since the former is already done. =D)

Since protein fold phylogenies gives some idea how the regulatory network evolved, at least on a tropic level, one could possibly backtrack towards simpler cells.
3 / 5 (2) Jul 21, 2012
The article specifically states they want to model all molecular interactions of the cells life cycle, not that of the entire cell. There is no point in modeling H20-H20 interactions, metals dissolved in the cystol, cell membrane dynamics (relative to lipid->lipid interactions), etc, which would encompass all the molecules of a cell.

I expected to hear about interaction between every molecule. It is not. It is a model of major interactions or of major processes at molecular scale

The Data S1 PDF goes into a lot more detail of what they actually modeled, and it is by no means just major processes. When they say complete computation, they mean it. They model all relative biological molecular interactions for one growth cycle. For example, they model in detail the interactions between nucleotides and associated proteins down to the different regions of the molecules.

Modeling of all cellular molecules would require super, or possibly quantum, computers
3 / 5 (2) Jul 21, 2012
The title is a little bit misleading. This isn't the first whole organism model. In the early 2000s the Japanese build a similar model using ECell, there are pretty good whole red blood cell models and finally Palsson pioneered the developed of genome-scale models comprising of sometimes 1000s of components which could do useful predictions. Other than than there are thousands of smaller models, go to biomodels repository to see the list of published models. I think what might be unique about this one is that is was trained against a large data set. However just fitting a model to data doesn't ensure that the model is good at prediction.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.