At first glance, the creature known as Caenorhabditis elegans—commonly referred to as C. elegans, a type of roundworm—seems remarkably simple; it is comprised of only 959 cells and approximately 302 neurons. In contrast, the human body contains somewhere around 100 trillion cells and about 100 billion neurons in the brain. Yet decoding the genome for this worm and digitally reproducing it—something that could spur enormous advances in the understanding of life and how organisms work—is a challenge for the ages.
“The project will take years to complete. It involves enormous time and resources,” says Stephen Larson, project coordinator for the OpenWorm Foundation. Larson, a neuro-scientist who is CEO of data software firm MetaCell, is not the only person focused on digitally reproducing life, or replicating evolution inside a computer. Researchers from a variety of fields are now attempting to decode worms, fly brains, and evolutionary processes in order to create virtual organisms and simulations of living creatures. It is safe to say that the field of executable biology—constructing computational models of biological systems—is coming to life.
Figure. Simulation of electrical activity in a “virtual brain slice” formed from seven unitary digital reconstructions of neocortical microcircuits.
Ultimately, this research could lead to a far greater understanding of how neurons fire and brains and entire organisms function. This knowledge would likely lead to new therapies and drugs for treating sickness and disease, but it could also produce new biofuels, as well as entirely new computing frameworks. It also raises questions about what constitutes life and whether living things can be engineered inside a computer.
Says Larson: “Understanding how organisms function would unlock many of the secrets of nature and change the way we view and interact with the world.”
Beyond Biology
The idea of developing virtual organisms and simulating physical systems through computing is nothing new. In the late 1940s, John von Neumann began exploring the concept of creating a computer virus modeled after a biological virus; he eventually developed the first self-replicating program. In 1970, mathematician John Horton Conway introduced a cellular automation system called Conway’s Game of Life, in which a person configures a set of circles, then the computer embarks on a rudimentary evolutionary process.
By the 1990s, a number of researchers had begun to explore the idea of producing digital representations of biological creatures.
For scientists, the idea of creating virtual life and artificial worlds inside a computer is rooted in practicality: it makes possible the study of the genetic information of an organism, or the creation of a virtual space to study how evolution and adaption take place. “Researchers can run thousands and thousands of replicates simultaneously. Every computer is essentially a Petri dish,” explains Christoph Adami, professor of microbiology and molecular genetics at Michigan State University. This approach also allows researchers to isolate specific components, including genetic coding, and “very carefully tease apart the different elements that go into the evolutionary process,” he says.
OpenWorm is an example of how this new frontier of biology and computing is unfolding. So far, “several hundred people” have contributed to the project in some way. This includes computer scientists, mathematicians, biologists, and experts in neuroscience. Among the core participants are more than a dozen academic and research luminaries, including C. elegans biologists Sreekanth Chalasani at the Salk Institute, Michael Francis at the University of Massachusetts Medical School, William Schafer at University of Cambridge, and Andrew Leifer at Princeton University. In addition, the organization has received computing input from the likes of Netta Cohen at Leeds University and Christian Grove at the California Institute of Technology (CalTech).
The OpenWorm project is now nearly seven years old. Larson estimates the project is 80% of the way toward achieving its first goal: assembling a digital model of the worm that allows researchers to simulate movement through simulated viscous fluids. The team hopes to achieve this milestone by late this year. This has involved mapping cells and functions in the worm’s body, developing software to run simulations, building a digital model of C. elegans, and constructing an algorithm that simulates the worm’s muscle movements—including how electrical signals travel through its brain and nervous system.
Mountains of data produce incremental gains, and coordinating all the research groups and silos is a complex endeavor.
So far, researchers at Caltech have developed the OpenWorm Browser, which relies on a Web or iOS interface to display a three-dimensional anatomical model and actions for C. elegans. The browser displays different layers, including the skin, alimentary system, nervous system, reproductive system, and body wall muscles. In addition, a program called Sibernetic uses a C++ algorithm to model and simulate contractile matter and membranes within the muscle tissue of the C. elegans. Another platform, Gepetto, provides an open source Web-based neuroscience simulation and visualization environment that simulates complex biological systems and their surrounding environment using multiple algorithms.
Not surprisingly, the data processing challenges related to OpenWorm and developing a life-like digital model are enormous; the overall task of understanding things like synthesis, reproduction, and digestion will likely take several more years. For now, researchers rely on a combination of classical mathematical and analytics tools along with machine learning to decode functions at the level of ion channels and cells. “Cells have a lot of extra machinery in them that is difficult to detect, and many of these activities and processes are completely ignored by artificial neural nets,” Larson says. Simply put: mountains of data produce incremental gains, and coordinating all the research groups and silos is a complex endeavor. Ultimately, “We may need to get to a new type of computing process to understand the exotic dynamics of natural neural systems,” he says.
Cracking the Code
The OpenWorm project is one of several current attempts to unravel the mysteries of living things.
For example, Virtual Fly Brain—a joint effort involving the University of Edinburgh, University of Cambridge, MRC Laboratory of Molecular Biology, Cambridge, and the European Bioinformatics Institute—is mapping the physiology of the household fly.
In 2012, Jonathan Karr at the Institute for Genomics & Multiscale Biology Institute at the Mt. Sinai School of Medicine in New York City assembled the first whole-cell model of Mycoplasma genitalium, a pathogenic bacterium that resides in humans. The model succeeded in predicting the viability of cells after genetic mutations.
In 2016, Stanford University bioengineering professor Markus W. Covert and a team of researchers developed a whole-cell computational model; they used detailed information from more than 900 scientific journals to gain insights into previously unobserved cellular behaviors.
There’s also the work of Henry Markram, a professor of neuroscience at the École Polytechnique Fédérale de Lausanne in Switzerland, director of that institution’s Laboratory of Neural Microcircuitry, and founder and director of the Swiss Blue Brain Project national brain initiative. His research has focused on synaptic plasticity and the microcircuitry of the neocortex. In 2005, he launched the initiative in order to reconstruct and simulate the mammalian brain, starting with the rodent neocortical column. Markram and fellow researchers are now attempting to reverse-engineer the circuitry of the brain—something that could radically redefine health and medicine.
Make no mistake, these projects extend far beyond a basic understanding of physiological mechanisms. An organism’s behavior is affected by numerous factors, ranging from its environment to its genetics. This means that even when scientists decode the genome of a creature such as C. elegans, it remains incredibly challenging to understand how cells function alone and together, and how they interact with the environment to adapt, adjust, and evolve. “Achieving a complete understanding of a worm requires incredible resources. Understanding the mechanisms in more advanced lifeforms is still very far off into the future,” says Herbert Sauro, associate professor of bioengineering at the University of Washington.
“It’s painstaking and arduous work to put all the pieces together,” says Alexander Hoffman, professor of immunology and microbiology at the University of California, Los Angeles. “It’s necessary to pull together research from a very large pool of existing literature, code all the information in a set of equations and parameters, and then work with computing software to relate model simulations to all the data. The problem for now is that there’s often not enough existing knowledge to deliver an accurate simulation of a phenotype—and so you wind up with gaps in knowledge that require further experimentation.”
Mind Games
A primary goal of these projects, and executable biology in general, is to produce reliable computer models that ultimately can be used to understand the behavior of cancer cells and address other debilitating or life-threatening diseases, ranging from multiple sclerosis and amyotrophic lateral sclerosis to heart disease and arthritis, says Sauro. “Understanding the internal machinery of cells and how they successfully orchestrate cellular remodeling in a way that doesn’t harm them could lead to faster and better ways to develop therapies and drugs.”
A deeper understanding of cellular activity also could help researchers engineer and reengineer organisms to produce biofuels and other chemical substances, or to produce entirely new categories and types of antibiotics and other medicines.
This field of research may also have enormous implications for computing, Larson says. Today, computational neuroscience attempts to faithfully reproduce the activity of neurons. However, neural nets do not exactly replicate the actions and behaviors of biological cells and neurons. “It’s not obvious what information processing neurons are doing when you consider them as biological cells. There are more exotic dynamics taking place, but we cannot see them,” he says. However, “It may be possible to develop more advanced types of computing systems based on biology. It appears that there is more we can do with neural nets than we have been doing with deep learning.”
Of course, the ultimate questions for researchers and society are where will all of this lead, and how exactly do we define life? At some point, digital code could replicate biological code for entire creatures, and researchers in synthetic biology might compile code to engineer new types of organisms—or autonomous devices, such as robots, that use biological models to think. This may bring the world to the highly discussed state of “singularity,” where intelligence becomes increasingly non-biological and humans transcend their biological origins.
Concludes Sauro: “In the future, we could see very different definitions of life. Once you start creating and evolving life-like behaviors in computer code or in synthetic biological systems and then applying them to the physical world, it’s possible to produce a very different reality.”
Further Reading
Adami, C., Hintze, A., Edlund, J.A., Olson, R.S., Knoester, D.B., Schosau, J, Albantakis, L., Tehrani-Saleh, A., Kvam, P., Sheneman, L., Goldsby, H., and Bohm, C.
Markov Brains: A Technical Introduction. Sept. 17, 2017. https://arxiv.org/pdf/1709.05601.pdf.
Basak, S., Behar, M., and Hoffmann, A.
Lessons from Mathematically Modeling the NFkB Pathway. Immunological Reviews, 2012. 246, pp.221–38. PMID: 22435558 PMCID: PMC3343698.
Palyanov, A., Khayrulin, S., and Larson, S.D.
Application of smoothed particle hydrodynamics to modeling mechanisms of biological tissue. Advances in Engineering Software. 2016. 98, 1–11. doi: 10.1016/j.advengsoft.2016.03.002.
Misirli, G., Hallinan, J., Pocock, M., Lord, P., McLaughlin, J.A., Sauro, H., and Wipat, A.
Data Integration and Mining for Synthetic Biology Design. ACS Synthetic Biology, April 25, 2016. Vol. 5, Issue 10, pp. 1086–1097. http://eprints.keele.ac.uk/3637/.
Join the Discussion (0)
Become a Member or Sign In to Post a Comment