Okay, thank you very much. Thank you, Victor, and also thank the organizers for inviting

me to give a talk at this really beautiful setting and a really interesting idea of having discussion rather than presentation. But let me carry on anyway. So I'm Paul Fremont. I'm the co-director of the Imperial College Synthetic Biology Hub. I have a great interest in how synthetic biology could be translated into useful applications, but also, of course,

in the spirit of this meeting, synthetic biology is an approach to understanding biology, and I think that's what I think Francois was mentioning earlier on. So my lab works on a number of different areas. I'm not going to be presenting data from all of these areas, but I'm going to tell you a little bit about our work on

cell-free platforms and also some work that we've been doing more recently on pathway engineering in the spirit of what Victor did. And Victor gave a fantastic introduction to some of the nuances, if you like, of synthetic biology and how we might take it forward, harnessing biology explicitly. Okay, so, I mean, Victor mentioned a little

bit about quotes, and so I thought I'd just quickly put in a couple of quotes. I found this article in the British Medical Journal in 1910, and it's a sort of quote, and I'll just read a little bit. It says, All natural sciences follow the same process of evolution. They begin by the observation and classification of natural objects and

phenomena, and that's the descriptive stage, and I do feel often that biology is still in the descriptive stage. Then they attempt to resolve these phenomena to determine the cause of their production, and thus they become analytical. And I think what this gentleman was saying at the time was that the idea of biological synthesis, the idea of going synthetic, is a really interesting one, because biology has only reached the

first two stages of description, if you like, and phenomena. And in that he said the idea of biological synthesis is a bold one, and yet it is no novelty. It has a cropped-up and imaginative literature of all ages, but considered as a scientific possibility. Its conception is a very recent date, so that was written in 1910. So if you wind

forward, this is a very large quote, but it's again talking about synthetic biology, so this is the Polish genetist, who I'm afraid I can't actually pronounce the name, so I apologize, but this is a famous sort of quote, but this is Wachla Przewalski, I think that's who's there, and he's a Polish genetist who came up with the idea

in 1974 about, truly about synthesis and synthetic biology, because he says, here, we will then devise new control elements, add these new modules to existing genomes and build up whole new genomes. I mean, this is essentially what we're doing now, and this was in 1974. And then he said this would be a field with the unlimited expansion

potential and hardly any limitation to building new, better control circuits, synthetic organisms, new, better mouse. I'm not concerned that we have run out of exciting, novel ideas in the synthetic biology in general. So he was very, very far ahead of his time. This

was only a few years after Asilomar and the introduction of genetic engineering and cloning, so this was a really interesting insight that he had of where the field would go. But then what is all the fuss about society and regulators and European unions? And this is some of the other literature, which we can't ignore, and that is the public literature.

This is what people read every day, and I grant you that some of this is from the UK press, which has got a particularly bad reputation, but things like scientists are cute at playing God after creating artificial life by making designer microbes from scratch, but could they wipe out humanity? Thank goodness they haven't yet, but anyway.

Reviving extinct species. This is from Friends of the Earth. Extreme genetic engineering in your ice cream. It's a very provocative picture, if you like, of a pipette and a lovely, beautiful ice cream. And then this is probably one that really was not very helpful,

which is brewing bad. Scientists find ways to cook up heroin at home. So this is creating quite a lot of fuss, and I think we need to be aware of, outside this wonderful place, that there are a lot of people that have got great concerns about synthetic biology. However, so what is all the fuss about? So synthetic biology, I think,

there are many definitions, and this is one definition which tries to capture what Victor was saying about the idea of building, designing, and constructing, and redesigning biological systems. And there are many reports. This is the most recent European operational definition, and there are some people, I think, in this room who were involved in this. I think it really

does capture very nicely what synthetic biology is. It's the application of science, technology, and engineering to facilitate and accelerate design, manufacture, and modification of genetic materials and living organisms. So we have definitions. So that's kind of where Victor came from. Now, just to put it in context, this is the number of publications in synthetic biology that name synthetic biology as part of the field. This is since 2010. You

can see it's rapidly rising. There are over 47,000 papers. This is the growth of the student competition called iGEM in synthetic biology, which I think there are teams from France and all over Europe and all over the world actually showing on this. So these are all young,

excited, enthusiastic researchers who are spending their time over the summer designing and building new biological systems. So you can see there's a huge growth there with almost 15,000 young people around the world have been through iGEM. And so it has this powerful vision,

if you like, for merging engineering design and practice and all of the associated tools involved in that, including obviously mathematics, computational modeling, and all the other what you call more hard sciences into the construction of biological systems and cells at the genetic and protein level. I think that vision is very persuasive to a lot of people.

So if we consider, I mean, Victor's already indicated this, but if we look at some of the very basics of engineering systems, clearly robustness and stability are key for engineered systems. And these are often achieved by these sort of four premises where one has systems control, one has redundancy obviously in engineered systems, one also has sort of

an idea of modular design, and also one has this idea of structural stability within the system you're designing. Now the question is, you know, how do we put that into context of biology? So we think about that. We can think, well, systems control, we have quite a bit of information and understanding about how biological systems regulate themselves. So we have control

circuits, we have feed forward, feedback loops, we have control networks, we have interaction networks. We also have redundancy. We have multiple genes that can carry out similar functions. We also have multiple regulatory pathways. If one pathway doesn't work, often another pathway will kick in. We also tend to have modular design. These evolutionally robust modules that

get passed from species to species, there's a functionality that's been solved and then it will be evolved or inherited by other species. That does happen in biology. And then we often have good structural stability, homeostasis. I mean, cells are incredibly good at regulating their

internal processes and life state if you like. So I suppose, and our hypothesis might be, are these features intrinsic to all complex systems, whether they're natural and artificial? And I think one aspect of systematic engineering, if you like, for biology

will clearly test that hypothesis. So I suppose the question is, can we learn about biology through design and construction? So, you know, biological systems can be considered as modular, I think. Functions primarily encoded in DNA, large knowledge of genome databases, large diversity of parts, if you like, increase understanding of molecular and cell biology

at all different resolution scales, new technologies to synthesize and assemble DNA, chemically synthesize, new computational tools to design and model, and obviously systems biology modeling and application comes into play here. However, I think it's important to realize that,

and I think everyone in this, biologists in their room should know, and hopefully everyone knows that, you know, there's some real challenges for engineering biology. And one is context dependency. So the idea that genes will function similarly depending on where they are within the genome is not correct. Evolution, adaptation, and natural selection, these are

very strong processes. This is that will change biological systems depending on their environment. Non-predictive stochastic behavior, which is part of the evolutionary process, if you like, self-assembly and immersion properties, non-linear dynamical processes, and multi-scale interactions. These are massive, massive challenges. And if you really boil it down, I mean,

living cells are essentially constrained volumes and very high concentrations of biochemical components. I mean, that's it. And so therefore, you know, biology is not plug and play. You cannot take one component, put it in the context, and assume that it will predictably function as you predict. This is not true, and it really poses problems. And illustrated here

is just a sort of network map for a really important eukaryotic mammalian signaling protein called mTOR, which is a PI3 kinase, which has functions in many different aspects of a

mammalian cell, including growth, including all sorts of functions within the cell itself. And I think, you know, this is a beautiful paper, by the way, showing the interconnectivities within a mammalian system that does provide a huge challenge if one wants to start engineering a part of that system or re-perturb that system. This is also a major protein

involved in cancer. So as Victor said, one approach may be to overcome this kind of, you know, almost overwhelming sense of complexity and bewilderment might be to try and develop

some sort of systematic design process. And I think that's what synthetic biology is trying to do, is trying to build things in a sort of more systematic engineering process. So using things like modularization, so interchangeable parts, interchangeable modules, using things like standardization, can we standardize measurements, tools, or processes?

And then using this idea of abstraction, which engineers use very successfully to try and sort of deconvolute complexity in some ways, to try and sort of allow people to cope with complexity. And systematic design aims to achieve ultimately robustness and reproducibility. But as I said, these are huge challenges in biological systems.

So this is already shown by Victor, and I think just to re-emphasize, it's a conceptual framework. It's not a literal framework. And it allows one to start thinking about biology at the genetic level as essentially functional genetic elements. And therefore, by building, you know, repositories and understanding of these parts and putting these parts together

in human-defined ways, just a simple transcription module promoter arrives in binding sequence, a protein, and a terminator. That's a module. And one could consider that module to be exemplified and analyzed and whatever and characterized. That module could become part of

this idea of going from parts to devices and then to systems. And I think this idea of abstraction hierarchy is actually a very powerful conceptual framework that allows one to start addressing this huge complexity that we're trying to deal with. This then leads on to this very slightly simplistic, if you like, but very effective design cycle where one can start doing systematic design, building, testing, learning.

And of course, the key aspect here is metrology, modeling, and sorry, metrology, data analysis, and modeling, and obviously learning about that process. And these are key elements of this design cycle. And the design cycle, again, is a framework.

It's not meant to be a literal thing. It doesn't mean you can build biological systems without doing this. But I think if you want to develop a systematic framework and learn about how you build biological systems, this idea of going through this systematic process is extraordinarily powerful and very useful. So I guess the big challenge is, and I'm sure

this is going to cause a lot of discussion over the next few days, is can we build new biological systems with standardized DNA parts? And already we are building registries and repositories of parts with nomenclature that people can use and analyze both digitally and functionally. Now, what about standards? So Victor led a beautiful project, actually,

called ST-FLOW, a four-year European Union project on standards in synthetic biology, which is incredibly, incredibly useful, bringing people together from all over Europe to look at standards. And I guess this is just a very simple standard. This is the first sort of thread standard by Joseph Whitworth, 1841. You can imagine how much impact the introduction

of a standard and a screw and a nut had on the world at that time. It was a hugely important development. And there are many other standards. Now, I won't go through all these. These are sort of the standards, if you like, that governments and consumers and businesses look for. But I think one key aspect is this idea of interoperability. And I think standards are directly linked to measurement. I think we need to understand

that can we standardize the construction of living matter? And this is a very big question. And I'm sure we can spend the rest of the week talking about that, just that one question. That is a huge question. And that is one of the, if you like, challenges and approaches that

synthetic biologists are trying to address. Now, the reason that we think that the systematic approach might be beneficial is because I think people realize that biological research, unlike my engineering colleagues' research or even chemistry and physics research, is often irreproducible. And my colleagues in physics and engineering find it extraordinary

that biologists actually live with this irreproducibility and can cope with it. But we do. And this is part of our descriptive storytelling, if you like, which we do very, successfully. Not all, but certainly we do quite a lot of that. And I think it's clear that biological data can suffer from irreproducibility. Now, I think the reason for that

is more a lack of technical standards and more a lack of sort of people doing the same thing constantly, using the same processes, using the same measurement tools, using the same strains and learning about the variability within systems. So from the standards consortium and our own thinking, I think you can think of standards as being possibly physical standards,

DNA standards, possibly functional standards, you know, standard measurement conditions, standard culture conditions, standard strains, i.e. I'm using the same strain as Victor's using in Spain and we use in London, and sharing data, standard strains, and then, of course, standards within digital information so that we can share all of this information digitally. And I think these are really important. Now, I do want to point out that synthetic

biologists do think that, you know, we all think we're really the new boys on the block. And, you know, this is a very good cartoon. I don't think it'll work. Let's do something different, something smarter, something cooler. And those kind of attributes do fit quite nicely with the synthetic biology. So I think we need to think about systems biology.

And the systems biology community have been going through exactly the same thinking that we are now approaching. And I think there is some overlap here that we need to bring into play and try and integrate both systems biology thinking and synthetic biology thinking. So what do we measure, if you like? What would you measure in a biological

system? There are many different things we can measure. And no one really fully understands what we could measure everything, not quite everything, but pretty much everything. So in a biological system, you could measure, you know, RNA transcripts quantitatively, proteins quantitatively. You can measure metabolites, lipids, glycans. You can get handles on post-translational modifications. Functional states are complicated,

you know, epigenetic state, growth state, noise within biological systems. You can measure noise, spatial localization, protein-protein interaction networks, regulatory networks, trying to bridge between genetic space, protein space, and metabolite space. These are complicated areas that one can try and develop models and try and develop understanding. So we can measure

quite a lot using the omics technologies that we have now. But no one clearly knows yet, I don't think, what we need to measure to really improve our sort of design, robustness, our design cycle. So this is where you kind of get this sort of

synthetic biology field going. There's this idea of a whole bunch of foundational technologies. You could reduce synthetic biology down to that, if you like, a whole bunch of synthetic biology technologies, which are things like design tools, you know, to build new genetic circuits, the synthesis and assembly of DNA, the parts and device characterization, and the standardized

measurements, and the whole kind of really meticulous measurement of your system and what's going wrong, what's working, followed with this very persuasive technology, CRISPR-Cas, genome editing screens, again, using the optimization of biology as a way to understand the design, if you like, cycle. And then, of course, working on how do we characterize,

what is a sort of, you know, do we have standardized strains, will we ever have standardized strains, can we work towards some sort of standardized host strain. So these foundational technologies, the idea is that they would fill into different applications. And, of course, the applications that people are very interested in now are shown on here.

These are not by no means complete, but there are a lot of work on foundational tools, therapeutics, novel drug delivery systems, agriscience, fine specialty chemicals, biomanufacturing, commodity chemicals and biomaterials, to mention but a few. Now, of course, this has been the area of industrial biotechnology for many years, so

Synthetic Biology is going to try and provide a new toolkit, if you like, to address some of those issues. So what are the current research trends? So when I look through all that literature I showed you earlier, these are the kind of things that were coming out from the current Synthetic Biology literature. There is quite a lot of people working on refactoring and redesigning, genome editing, genome construction,

automation standards and tools, and then some sort of quite a bit of literature, also in some of the social sciences, but open source and descaling. There's quite a lot of work on that. There is a growing interest and excitement in the idea of creating alternative biological systems using exobiology, XNA, artificial cells and cell-free systems,

this idea of building cells from the bottom up. And I think this is an area which is actually very, very interesting, and there is some kind of interesting integration of cell-free systems and the kind of alternative biological systems and what I would call more the mainstream Synthetic Biology. So that leads me nicely onto work that we've been doing on

cell-free systems, and I'm going to now just switch gear a little bit to what we've been doing on cell-free systems. So cell-free systems are really interesting because they are essentially the cell extract with the membrane peeled off, and all of the ingredients within the cell extracted the genomic DNA removed, and essentially it contains ribosomes,

some membrane vesicles, and some cellular proteins. So it's a crude sort of lysate, if you like, from a cell, but it has the great ability to be able to translate and transcribe DNA within as a biochemical reaction and assay. So it takes out the life of a cell,

if you like, but uses all the ingredients within the cell to carry out reactions. So this is a sort of scaled-down version, if you like. So what's interesting about cell-free systems is that you can use part of the glycolytic pathway, which is the ATP-generating pathway that exists in cells, and also the TCA cycle is existing within cell-free exteriors, but you

can provide new energy sources. So one common energy source is 3-phosphoglycerate, but the people are working on cheaper energy sources. It's clear within cell-free systems you have components of oxidative phosphorylation, so you do have the ability to create ATP within the system, although you do need to add an ATP-regenerating system. You also need to add amino acids and

other essential cofactors to allow the system to kick off. But the point is that within that system, you can get transcription and translation working quite reproducibly and robustly. Now, there is an alternative system, which is the Pure system, which was a beautiful system, essentially first published by some Japanese colleagues,

that went to the effort of purifying all of the machinery of transcription and translation and re-continuing that in a test tube. If you like, that's the sort of, you know, not only is it pure, it's sort of, you know, beautiful science, if you like, of reconstructing the basic

components that would allow transcription and translation to occur in a test tube. The disadvantage of the Pure system is it's extremely expensive and awfully difficult to get running routinely in the lab, but it is a beautiful system nonetheless. So, what are the advantages of cell-free systems? So, you can do transcription and translation,

you can do DNA circuit prototyping, you can use them for biosensors, environmental testing, we've got some projects on that. You can actually do enzyme pathways for fine chemical and drug production, you can make recombinant proteins, and you can do toxic pathways. It is scalable, you could scale it up to 1000 litres as shown by Sutro Biopharmaceuticals,

but it is probably best if you're thinking about producing products that are high-value, low-volume products. Now, the metabolism is simple and cheap, and it's easy to modify, so you can do all sorts of interesting things within cell-free systems. So, as a test bed, it's a very useful system to operate. And in the context of what I described

earlier in synthetic biology, if you were prototyping parts, a few years ago we had this idea, well, if you've got all these parts, you want to measure the functions or the quantitation of a promoter's or rhombus and binary sequences or whatever. You know, to do this using standard molecular biology, it's sort of a long process, and we wanted to speed that up and see if we could explore whether the information we got

from in vitro systems was very similar to the information we got from in vivo systems. So, we set about doing that, as if we could use it as a prototyping. And this is the idea of taking parts, doing all of the molecular biology, you know, the ligations, the transformations, the liquid cultures, the growing, the measurement.

It's a very tedious process. So, the idea was, if we want to do, if synthetic biology is going to become this kind of engineering field, you want to have thousands of parts characterized to some level of quantitation so that you can inform the various modeling aspects of the field. So, we decided a few years ago, James Chappell, a Ph.D. student,

to look at that in detail. So, we took a bunch of parts, a bunch of promoters, a bunch of rhombus and binary sequences, and we hooked them all up to a GFP reporter, and we did a very, very simple experiment. We measured the GFP, if you like, production in vivo, in a steady-state expression system, midlog.

And then, using BL21DE3, using M9 minimal media, 30 degrees, we took the same cell-free extract from BL21's DE3, 30 degrees, but obviously, it's a completely different reaction, and we measured the production of GFP from the same parts in the same context in both systems.

Now, to our surprise, we found on this side that the measurements of GFP, the relative measurements, the relative production of GFP between in vivo and in vitro were similar. We were quite surprised, obviously significant error, there are some sort of largest error bars,

but the relative strengths of some of the promoters shown over here, you can see that, you know, in vivo is in the gray and in vitro is in the white, and you're getting kind of nice, relatively good correlations between in vivo and in vitro, and it was the first

time we'd, I was unexpected. And we also did some reduce with rotors, and we got similar data, and we published this, so I won't spend too much time. At the same time, a whole bunch of other papers came out as well, and there was this sort of acceptance, if you like, or not, proposition, sorry, proposition that said that for simple DNA regulatory parts,

the ones that had been studied, they showed similar kind of functionality, similar quantitative behaviors in vitro and in vivo, which was quite surprising. However, as all biology shows you, this is not the case. So now, this is a library screen of promoters we've been doing recently, and we found some really quite extraordinarily strong

library. This is an in vitro, this is an in vivo screen, and we found, sorry, an in vitro screen, and we found some really, really strong promoters, and then no need to look at the data, but it was an extremely strong promoter. And it turned out that, and here it is here, this is the normal Kelly promoter down here, and this is the promoter we found. It's a really unexpected observation as we were, again, this

descriptive nature of biology. As we went through all of these different sequences, we found this extraordinarily high promoter. And I think what was interesting about it, as shown here, just shown here, when we went to look in vivo, we could not replicate at all that promoter strength. It looks like it's the same promoter strength as the

Kelly promoter down here in vivo. And, you know, there are also, and we're exploring why that is. Subsequent to that, Zak San and others came up and said, well, actually, this does break down. So this idea of in vivo in vitro does break down.

So I guess the way I could pitch that would be, well, could we, you know, could we use this in vitro in vivo kind of comparison as a way to tell us about context dependency? And I think that's something we're going to explore with this very, very high producer, this promoter, which is essentially two base changes, which is quite extraordinary, and we need to work out why that is. So then we're making cell-free

extracts from different cells. We're going to explore cell-free extracts as a platform. We're going to try and compare them from E. coli, different strains of E. coli. So this is MG, this is Rosetta, BL21s, looking to see if we can learn about any of the sort of

phenotypic functionalities of cell-free extracts. This is now bacillus subtilis, which we've managed to get optimized, and we're going to be looking at bacillus subtilis as well. And then this is, that's the optimization of bacillus subtilis. And then we're going to be looking at bacillus megaterium, which is a very interesting organism, has been thought of as a very important organism potentially for an industrial

production setting, and we're doing this in collaboration with colleagues at the branch flag technical university. And so we've made a cell-free extract from bacillus megaterium, and we're getting extremely good production of proteins within the bacillus megaterium. Now, in the context of that approach, we're also developing some real-time

messenger RNA measurements, and the idea here is to try and provide quantitative data that would allow you to assess the cell-free system in a more quantitative mathematical way, or modeling way. And we are getting very nice data showing you get very nice births and decays of messenger RNA. We're also looking at trying to do very,

very high throughput analysis. So this is on our echoliquid handler. So this is 108 conditions in triplicate, three times DNA, six times repressed juices, and we've been developing a model. So one of my senior researchers in my lab is a physicist, actually, originally,

and he's been developing a mathematical model to try and develop. This is the model here. It's a Bayesian statistical inference model. It's around trying to map out the parameters that we don't know at all within our system what they might be. And the modeling parameters that we're interested in is polymerase binding, messenger RNA synthesis, messenger RNA degradation,

and then GFP synthesis, GFP maturation. So this is Jerry McDonald's work. And the idea here is that we can start doing simulations as well as experimental observations. And of course, here, we can start providing the quantitative details that would allow that model to become much more, not better, but sort of more informed, if you like, on experimental data.

And so I think cell-free systems, so here's the kind of summary of that, if you like. So cell-free systems, I think, are a very useful testbed to explore part of the design cycle of synthetic biology, but they're also a very good testbed to start

and develop slightly simplified models in a non-living system, but having all the central sort of broken-down parts of life, if like in terms of metabolism. So we've been developing a whole extract model here. This is James' work, doing experiments, metabolomics, trying to infill this model, and also doing proteomics as well, to build up a cell-free kind of scenario. And then we'll be comparing that with our

different cell extracts to see if this breaks down, depending on the particular extract. But it's a, yeah. Okay, so I mean, that leads on to the obvious question, I think, which a lot of people are interested in, which is could we build a cell from the bottom up using sort of subsystems, if you like, maybe using cell-free systems? And if you break down a cell, I think this is the idea of modularity,

functional modularity. And here you can see, if you break down a cell, a very simple bacterial cell, there are discrete components that you can think about. So there are actuation, sensing, so sensing, actuation, export, communications with the

outside regulation and computation within the cell, signaling, and metabolism. And I think one of the challenges, and one of the exciting challenges, I think, cell-free and synthetic biology, per se, can offer is to try and develop sort of these subsystems, so, you know, sort of to build these subsystems and see if they can work. Now, clearly,

subsystems that involve a membrane compartment will be difficult, but we can certainly start looking at regulation and computation or even some metabolic subsystems within the cell-free system, and that's one of the goals that we're going to be moving into in collaboration. Okay, so finally, I'm not sure how long I've got left.

Okay, so finally, I just want to go into some work we've been doing on pathway engineering here. Now, I think everyone realizes that cells can be used as metabolic factories, and that's been around a long time. It's been around so long that we forget that

industrial biotechnology has been with us for, you know, 5,000 years, maybe, or a few thousand years, or whatever, since we started making wine, I guess. But anyway, it's a very, very sophisticated industry that has been using cellular systems to produce and manufacture components and some really major pharmaceutical components as well. So I guess when you look at

industrial biotechnology, you can see that all of these products that we take kind of for granted, there are components within these products that are being or can be manufactured using biological systems. So it's very apt that we just had the climate change big convention in

Paris just the other day. Everyone's very excited about moving to this non-petroleum-based world that we're all going to have to live in, and clearly industrial biotechnology has a massively important role and synthetic biology to provide the components and chemical entities that we all need for all of these everyday life systems, or we adapt our lives to not

live with them, which is going to be difficult. So of course, scale-up is a huge problem in industrial biotech, and it still continues to be, and I think that is one of the big challenges I think that synthetic biology is going to have to try and address, because it's doing stuff in the lab. So the question is, can synthetic biology accelerate the construction

and prototyping of synthetic pathways for the production of products, if you like? And these are the kind of areas I think that are important for pathway engineering. So we have bioinformatics, clearly, flux modeling, combinatorial pathway assembly, metrology, in vitro and vivo, and chassis host cells, and then we need to go through this

testing and prototyping. So we've been doing some work on a new kind of Golden Gate-based E. coli kit for part assembly, which we're just about to publish, and the idea here is that we can accelerate the production of different pathways, different combinatorial circuits. This is based on

Golden Gate, and it's a sort of plasma kit that we can use for many different range of applications, and we put in all sorts of variants into the system. And the Golden Gate assembly strategy, a busy slide, is extremely powerful technology. It's been used around a lot. It's a fantastically systematic way of assembling multiple, multiple components. There are other methodologies for assembly, but I think Golden Gate gives you quite a lot of

combinatorial variations. So we've made a Golden Gate kit for E. coli, where we can start assembling these modules, which can then be assembled into pathways and into greater modules shown here. We've also put in some variants in the system to allow you,

when you're doing combinatorial assembly, that you can keep some of the components constant, and then just assemble certain parts of the system, which I think could be very useful. We put in various other things, like protein purification tags and all sorts of other things. So we're hoping to submit that to Addgene, so hopefully everyone will be able to access that and find it useful, like we find it useful.

So we've been thinking about products, and clearly there are a lot of interesting products that people are manufacturing and making, very, very complex products. But we wanted to develop more of the platform technology to allow us to see, how would we make a pathway? So we chose something called Raspberry Ketone.

I really like raspberries, and there is this sort of product here that comes out of raspberries, which essentially gives you the sort of essence of what raspberries are. Now, obviously, there are various economic factors, but we just wanted to look at this pathway. We're not really interested in that. We're more interested in using the pathway as a

simple prototyping testbed. So here we do our combinatorial DNA assembly, we do the cell-free prototyping, we do high throughput LC-MS, and then we bring in structural biology and other aspects into the process. So here's the pathway. It's quite a simple five-enzyme pathway, four-enzyme pathway here, and there's another enzyme involved here. It starts at tyrosine,

and it goes to the raspberry ketone through a series of enzymatic conversions. You can also come in from a chemical, 4-hydroxybenzaldehyde. You can do a chemical process to produce this hydroxybenzylacetone component, which can then go to form the raspberry ketone. So it's a sort of biochemical pathway, enzymatically catalyzed,

produces a natural component, so we were just using it as a testbed. So the first thing we did is we purified all the enzymes, and we did very pure in vitro biochemistry. So here are all the enzymes here, and all this has worked unpublished. So these are the four enzymes we purified. We then increased tyrosine. This

is the substrate concentration, and then we looked to see what we produced over time and in terms of the concentration of the products. And you can see here that just from this very simple piece, this is the enzymes themselves all together, put in substrate, get our product.

You can do very simple Michaelis-Menten modeling, all sort of modeling on this, and you find that actually there is some sort of product inhibition in this pathway, where you get this four-cumarate accumulating where the product as a function of increased tyrosine concentration. So very simple observation, but a very important observation

if you're trying to build a pathway in vivo. We then did some screening with different ribosome binding sequences using our Golden Gate EcoFlex system, and again in vivo, and these are just a series of ribosome binding variants and different promoters on different genes. We were finding different outputs from the production process. So combinatorially,

we were finding actually the product is in red, and that's the product we're trying to increase the yield of, clearly. And we were finding all sorts of interesting correlations in here between ribosome strength and different intermediate products and different products. And clearly, we're interested in trying to develop a more intrinsic understanding of that

in terms of a model. But in general terms, what we find is that particular combinations of promoters are producing different sort of points within the pathway where we're getting product getting caught up, we're getting product inhibition, we're getting all sorts of interesting and unexpected outcomes that we didn't know.

Then we decided to try and do some structural biology, and we built this model into a crystal structure where we're trying to change the requirement for NADPH to be NADH. So we've now got a very nice mutant here which can use NADH instead of NADBH as one of the components in the pathway. So I guess from this engineering kind of like pathway engineering

approach, I think what we're beginning to realize is that the landscape of data space that you need to explore in a four-enzyme pathway to try and optimize the product production, if that's your target function, is actually a very, very large space indeed. And there are

many, many different nuances and unexpected consequences of changing various parameters, which I think Victor alluded to in his very nice introductory slide. And clearly, you could start thinking about maybe applying weights, if you like, which is kind of what we're trying to do here. And so having a mathematical formulation around where

to go next, what to explore, what to change in this kind of design would be extremely useful. Okay. So I think from our initial sort of unpublished data so far, refactoring pathways requires, I think, multiple approaches. Promoter strengths we find are often inversely correlated to the production. So you think if

I have high production of different enzymes, I'm going to produce, but that clearly is not the case. And there is clearly a lot of unknowns, and I think that's been well understood in the literature. Cell-free assays, I think, have been helping us to make decision points along that reaction pathway. And if you wanted just to finish up now on the challenges moving forward

for the field, and this is more of a discussion point, I think, if we are going to become a kind of sort of more engineer-y type of field, I think we need to develop technical standards. We need to have sharing of parts. We need to have parts that are shared between multiple labs, multiple groups, ever in the world, openly, easily, so that we can learn from

all of the information that we need to make this process become much more systematic and predictable. We need to share, I think, detailed data on failures and successes. So the great thing about biology, and I'm not sure this is true in other sciences, but certainly in biology, we never share our failures. And to be honest, most time in

biological experiments don't work. Probably 90 percent of the time, if not more, they don't work, and we don't, and we ignore it. And so we need to start thinking about failure and sharing that and looking to see what works, what doesn't work. And we clearly need to integrate systems biology thinking and approaches to try and

really harness what's already been done in systems biology into synthetic biology, and those are just some of my own thoughts. So I've rushed through a lot of stuff there, but I just hope to give you a flavor of some of the things we're doing in the lab, and how Cell Free is producing to be a really extremely powerful technology.

And the work I described is Richard Kerwick, Jane McDonald, Simon Moore, and we're funded. And thank you very much for your attention.