starting with digital code and starting with DNA. It is software and it does build its own hardware. I think that's an interesting science question. What do we need? Do we need some ribosomes and TRNAs and lipids and can we get some cells bootup. I think that's going to be important about understanding origins of life. But we are in fact building upon 3.5 billion years of evolution. The argument is because we're using genes that we have discovered versus inventing a new gene is a spurious one. It's like they didn't create a new car because they built a battery from one source and they combined that to make the Tesla, a pretty exciting electric car. My team has discovered a variety of genes known to science. We're up to 40 million. There were less than 1 million when we started. These are going to be the future design components.
I think biobricks are an important teaching tool. It's great for getting students involved but the number of genes on this planet I'm sure will top out somewhere over 200 or 300 million. We're dealing with a lot of design components. Nobody is going to patent them. And I think combining those in new ways, now using these tools we had in the proof of context experiment is what the future of this field is going to be. We couldn't do any of this until we did the study that was just published in science. We were able to make these really large pieces of DNA. But until you can boot them up in a cell and get them activated, it was an interesting academic exercise. So it's very different from what's happened before in molecular biology. This is a new set of tools starting from a new vantage point. And as Drew said, it changes a lot of the rules. Scientists sort of controlled who got what, whether they sent them their cell line or their DNA clone for a gene. Now anybody who has access to the Internet, if that information is in the public databases, you can download it and you can make those genes. Now, any virus sequence that's in the public databases can be pretty readily remade. Fortunately, not all of them is the DNA ineffective. Smallpox is one of them where having the DNA from smallpox on its own can't just boot up readily. But these tools are there. It's a different startling point -- starting point. All you need is the digital information in a DNA synthesizer. So building the pieces of DNA was an interesting technological challenge. Getting booted up was straightforward biology and molecular biology. None of it is cloning. These are totally misuses of terms. Cloning means everything and anything to biologiesists and it's sort of a collecting term and making copies of cells. It's making copies of DNA. It's splicing DNA. So we think it's an irrelevant term for what we do. We use the term synthetic cell because every protein in the cell, all the constructions in the cell are derived from the synthetic DNA. The cell that we use as the recipient cell, all its characteristics are 100% gone after a few replications. So everything in the cell that we have is from that synthetic DNA and, therefore, we define it as a synthetic cell. It's a cell that never existed before. Of course, we use copies of existing genomes. I agree with the very statements we're very early on in our knowledge of biology, but we definitely have new tools now to get there. We're using some of these tools to make new vaccines so we have a program that is funded by the NIH to make synthetic components of every flu vaccine that we and others have ever sequenced. And we can recombine these and make a new flu vaccine candidate in less than 24 hours that we're working with no virus and it's very possible the flu vaccine you get next year will be from these synthetic DNA, synthetic genomic technologies. It was announced last year that we have a program with ExxonMobil to try and get cells to capture CO2 and make basically a biocrude that can go into refineries. We still have not found any cells that can do this naturally at the levels that are required. So at the very minimum, it's going to need extensive engineering, but I'm absolutely certain, at least by the time we get to version 2.0 of the cells, they will be completely synthetic as will most things going forward in an industrial environment. Definitions are important. The definitions can be found in our scientific publication. I think this is an area that Drew Endy's students show we are limited more by our imaginations now than any technological limitations. I think having an intelligent ethical framework for this new science to emerge in is absolutely critical. Thank you very much.
>> Thank you. We appreciate your views and clarification on the definition. And also impressing upon us the value of the technology coupling with the science. Dr. George Church is our next presenter. Professor of genetickic at Harvard med. He has spineered innovations in reading and writing DNA, he directs personal genomes.org with a goal of enabling open access integration of full genome sequences, environmental and trait data goal of working toward 100,000 individuals. Very interesting application. Again, this session being on applications, very eager. And Dr. George to hear what you have to say. Thanks for being here.

>> So thank you for the time here. As soon as my slides come up, I'm going to talk almost entirely about application. And it's going to different a little bit from previous talks in that I'm not going to talk about introductory definitions and in particular about what we can't do or having done but what we are doing. So this is my thank you slide. And my conflict of interest slide.
[LAUGHTER]
Next slide, please. Can I control this? There we go. Perfect. Okay. Sorry. And so as a graduate student, I work with Greg sutCliff, which we did a ridiculously high cost even though we were students, this has been used in many recombinant DNA efforts that some of them are listed here that we're really single gene efforts. What's wrong with this picture? This fellow is not using safety goggles. He's not properly grounded for electropouration. But the main thing is we've gone well beyond main genome engineering I had in the last slide. We have gone beyond minimal slides to these fast robust useful cells. We're focusing on lower costs. We have talked a lot about scaling up, but not lowering costs. I will focus on that. We look forward to whole genomes, but most of what I'll talk about is doing a little bit less than a whole genome but on a genome scale. And the question is, why do we do things on a genome scale? And then there's safety and security for the reason for doing things on a genome scale. And evolution is a unique capability that we have that most other fields in engineering do not have. And my major takeoff for all this is that we are going much faster than it appears. And we should not be reassured that biology is not capable of engineering and there's no difference between what we're doing and what I did as a graduate student. Why, genome-wide? We need to know why. Genome engineering is a commonly used term and is also a couple of genes done in the chromosome rather than on a plasma. Big deal.
Metabolic you might have a pathway or small network. 30 genes or less. But genetic code offers us multi-virus resistance and safety measures and some use of new amino acid and this is genome wide and one of the few articulated goals that is genome wide. The safety component is incredibly important. This is not meant to just be an analogy or images. But we have interoperable parts. These are all from cars but the same thing applied in biological design. Cost effectiveness, standards and isolation. We need to -- is it sufficient to have a set of rules and guidelines if there isn't testing, if there isn't surveillance? You can do licensing like driver's license but you have to do surveillance to make sure people are obeying the laws. And then again evolution is something that's new. There have been recommendations in 2006 and the next slide 2007 which I think don't go far enough. We talk about preferred practices. We pragmatically talk about federal grantees and contractors. There's a lot more out there than federal grantees and contractors. The Sloan 2007 went a little further than this. But we need to have surveillance and enforcement. And so back to my earlier recommendations on really licensing the entire ecosystem in synthetic biology, it's important. We need to have surveillance and testing of systems that are proposed to go in. And this is not restricted to bacteria. We have a very active human synthetic biology community and human do-it-yourself community. Some of my undergraduates have gone and sequenced part of their genomes on their own without F.D.A. approval and without really using any special equipment. And this is a whole another subject we're not going to talk about. But do it yourself or do it ourself biology and bio weather map and so on. We have studied vaccinations. Genome engineering, some success stories, we already mentioned one but also propane from DuPont to a $400 million project 90% successful. It only involved eight foreign genes plus 13 -- I'm sorry 13 down and six up regulations in the e.coli genome. 27 changes was a lot of work back then. I'm going to talk about hundreds of changes that we have incorporated. These are two other companies that I helped start that are not in the future but are already making thousands of liters of production scale fuels, either from biomass or from carbon dioxide and light. These are making alcanes, diesel and gasoline. Part of this is the success of comparative genomics. You can look through the bacteria for those that make trace amounts, Greg sort of alluded to trace amounts of the alcanes by taking fatty acids and reducing and decarbonnallating them. You can look at the genomes that produced and those that didn't produce these trace amounts and then you can identify the genes and overproduce them. Rob Carlson alluded to this exponential curve. This is actually quite different than his curves, although basically the same. What's different is that around 2004 or 2005, there was an increase in the rate of this exponential curve from 1.5 to 10 fold. And more importantly, this is a gap between the -- thank you. Between our recent huge increase in second generation or next generation sequencing and synthesis. And we're still stuck in the first generation for gene synthesis in the companies and genome synthesis that we're using first generation sequencing and synthesis for the most part.
There are 21 next generation sequencing technologies and 21 companies that go with it. And I am an advisor for about 16 of them. And similarly, there's a next generation synthesis off of chips that we've been doing since around 2004. This has lagged a little bit behind from making agains and genomes but it's certainly terrific for making short constructs.
Working in the cells, it's one thing to make DNA but getting the work in the cells, there are many tools. These are protein based specificity tools. And more general tools which are DNA based, they don't require specific proteins to put it in precise locations in the genome to make precise changes. But some of these involve single stranded DNA number 3 and number 4 in particular. And we have automated this in order to bring down the cost and extend our capabilities industrially. One of these is called -- or the general term is multiplex genome engineering or MAGE. And this has one particular implementation shown on this slide but there are many others. You can see it's a catch-all phrase. This one uses single strand nucleotides that use computer aided design to optimize secondary structures, optimize the position and length. You have to have a mismatch repair turned off for some of these. And there's a special proteins. But the key point is in a few years, we move from an efficiency around 10 to minus 4, 1 in 10,000 to 25% to 100%. And now we can get up to 8 mutations per two-hour cycle and we can just continue the cycle, 8 changes precisely in the genome wherever you want. And I'm sorry. You can make up to 1 billion different changes in a population. I'll show you an example where we did 100,000. This is Harris' prototype. A computer aided design of the upgrade. This is the sphul upgrade. This is applying it where we made 100,000 genomes, not one by one, but in a mixture. And it shows the awesome power of accelerated evolution in the laboratory, where we could make these 100,000 genomes focusing all of the changes in the known path ways, including putting in some genes from other organisms. And in three days, we can get the highest yields we have ever seen for this hydrocarbon lycopene which makes tomatoes red involved on the order of 24 genes. Another project that we have done which is less commonna torial and allows new amino acids and has safety features, here we changed all of the codons into TAA genome wide in order to free up that codon and allow us to delete the cellular factor that recognizes it. This can be generalized. There are 64 codons of these triplets and we have targeted nine of them. This allows us to do three things. New amino acids, safety features and multi-virus resistant which itself is a safety feature. We have these nine. We have done one of these nine codons that we're targeting out of 64. We have synthesized all the DNA to do the remaining eight, at least proof of concept on the essential genes. And another topic that is far beyond what we can talk about today probably is the project where we're making ribosomes and Greg aleaded to an in-vitro system which has interesting commercial applications. The key thing here is just changing these nine codons would require changing just 2.7% of the genome, not the whole genome. But if we're making these optimal 90, we have compiled the genome two and a half fold over and we essentially have remade the genome, even though we've only changed 2.7% of it. And that lies in the future, and it remains to be seen which is more efficient. Doing it all synthesis all at once where we'll probably have multiple failures or doing it one at a time. And just as a quick last slide or two is this issue of safety in terms of isolation. You can have physical isolation or you can have biological isolation. The changes of the genetic code, the genes can neither go out or come in that are functioning. The critics of the genetically manufactured organisms have wanted it.
A third way that it's isolated is physical and genetic and it's this metabolic dating back to the early days of recombinant DNA there was this acid that was used by deleting the biosynthetic pathway that you made the bacterium dependent upon that. It's not common in the environment, but it does occur. And that's one of the down sides. Some of these other SACB or tox-antitox pairs are used but as counter selections. But they are ways of having the cell self-destruct but they have the problem that they can be lost just before you need them. So they are not ideal. So we think going forward using the new genetic code to allow us to design multiple essential genes to have multiple dependencies that have been used in Peter Schultz's group. So in conclusion, just to remind you, you know, where we think we need genome engineering and synthetic biology, it's in making biology safer than it already is and this involves really using some of the lessons of other engineering disciplines, interoperable parts, hierarchyial designs, cost effectiveness, standards, isolation, testing, redundant systems, surveillance very important, not just surveillance of government grantees. Licensing at every part of the ecosystem. And focusing on this ability to evolve both in the lab and outside the lab. Thank you.
>> George, thank you for that. Your message is loud and clear in the face of advancement and technology advancement is astounding. And some near-term applications are very exciting. And also clarifies and I appreciate your last slide. And it was used before to help clarify for us what some engineering challenges are going forward. Our final speaker in this panel is Kristala Jones Prather. Dr. Prather is an Assistant Professor of chemical engineering at MIT and worked in industry as well as academia. Has been recognized for her work with numerous awards and investments. She is a research young investigator and received technology reviews TR35 young investigator award. She has also the NSF investing in her through an NSF career award. She's an investigator in the multiple institutional synthetic biology engineering research center funded by NSF. Welcome, we're pleased to have you here.
>> Thank you very much. Let me start