Week 6: Darwin on Steroids
Bio design, diversity & selection
This class was in the middle of science and fantasy, with George Church answering hiw own question "How do we compete with Darwin?", referring to the vast amount of time and space for which genetic code has been allowed to evolve through on its own. We discussed directed evolution and the idea of a “Human Genome Project 2.0”, where instead of reading DNA we are synthesizing and writing it!
1. If humanity were to undertake a Genome Writing Project, what would be the benefits? What types of new science and engineering would be enabled if we had such a synthetic human genome? Please provide specific examples.
The ability to design and assemble a functional human genome opens a whole new dimension to the field of synthetic biology. One that intrigues me the most is the notion of the word "human" as an Earth inhabitant. Our genome has been optimized for ~3 billion years to be able to survive on Earth and be evolutionary competent against other species and earthy nature itself. So, if we have the ability to programmably and rationally alter our genome, could we make it optimal for life on another planet?
As a specific example, we can look at the extremophiles: organisms that thrive in extreme environments, under high pressure and temperature and even lethal dosages of radiation. If we had the ability to copy, recode and paste the genes that make those bacteria competent in such fierce environmental conditions into an synthetic human genome, then we could engineer a true Martian, which unlike Matt Damon would be able to survive on the real planet Mars! An inspiring talk on this topic was given by Lisa Nip from Molecular Machines group at Media Lab, which you can see here.
2. Conversely, why might we not want to proceed with such an endeavor? What are the risks?
As this synthetic biology project directly affects human life, there are many risks involved in such an endeavor and a set of ethical questions are rising. The one risk I find more challenging, is the definition and handling of an "error". Assume that we assembled a genome and started growing a human out of it, but in the middle of the process we find out that it has an error that will lead to premature death, uncurable cancers, or mental disorders. If even writing the previous sentence felt cruel and inhumane, would we ever have the ethical prowess to experiment with a human life ?
3. Map out a technical strategy for synthesizing a human genome. What technologies would be required? What are existing tools we could leverage? For certain tools that do not exist, what should their capabilities be?
Let's first be clear about the magnitude of such a synthesis project: The human genome has a size of about 3 billion base pairs, organized hyper-efficiently in 23 chromosomes and optimized for copying itself with minimal errors. Thus, if we want to really synthesize a human genome, we should think not only of how to assemble it, but also how to store it efficiently inside the nucleus of a human cell. While the following answer focuses on the synthesis problem, as you can see in Fig. 2, the DNA is packaged in a very complex way in order to take up as little space as possible inside the nucleus. It is very challenging to replicate this packaging by creating synthetic nucleosomes and will there exist publications towards this direction, generating synthetic chromatin structure is still an open, interesting research question.
I believe a good strategy towards this project is to synthesize individual chromosomes, which has been done already as you can read in the fascinating article from Ewen Callawy in Nature News. In that project undergraduates managed to synthesize a DNA molecule of around 270,000 bases, from smaller fragments with a Gibbson assembly procedure. In order to make this happen in a larger scale though, we really have to push the limits of DNA synthesis so that we have to perform only a few steps of Gibbson assebly or other stochiometric procedures. As Prof. Joe Jacobson said two classes ago, we really need a technology to be able to synthesize large DNA molecules with as few errors as possible, and we need that on-chip so that we can do it fast and high-throughput. So while we can utilize current gene synthesis tools (photo-electro-chemical) and assembly techniques (check here for an overview) we really cannot go much further that a few hundreds of thousands of base pairs.
Thus, we require a novel DNA synthesis tool, with capabilities in the order of billions of base pairs. I imagine an array of DNA synthesis chips that produce genes in the order of one to five thousand base pairs, subsequently fed to a microfluidic device multiplexing a series of Gibbson assebmly steps in a hierarchical way. In the middle of the assembly line, we can have high-throughput sequencing steps in order to validate the synthesis of smaller fragments and also amplify them for the subsequent step of synthesis. An inspiring paper along this direction is from David Kong et al. in which they fabricate a high-throughput, accurate, inexpensive microfluidic gene synthesizer (see Fig. 3).