So as as. Bob just said I'd like to thank you for coming out on this cold and blustery morning straight out of a Winnie the Pooh story. So I've spent. I just completed my first year as a faculty member here at Georgia Tech. And that this that first year has really been a wonderful experience for me. In part because Georgia Tech is remarkably multi-disciplinary supports multi-disciplinary research and as you can see here I am. So this is my primary appointment here in the school of biology but I'm also an adjunct member in the school of chemistry a member of Id be as well as a member very active member of a number of centers these are just trivial membership your paper memberships. These are centers whose goals directly align with research teams in my laboratory. And so the this multi-disciplinary supportive atmosphere is is certainly reflective also in the composition of members of my research group where I have nine people currently and they're all doing different things. So I've molecular evolutionists chemists molecular biologists computational scientists. And. Globally. They're all contributing to a common theme but individually they're all doing. Separate things that don't have much overlap. In case so globally what we're trying to do is develop something that we have termed in evolutionary synthetic biology. And we want this to be more than just a simple catchphrase we really are trying to turn this into a paradigm so to speak. A way to do research where on the one hand we want to use evolutionary. Suppose the models to develop engineer proteins as well as potentially engineered organisms. On the other hand we want to generate synthetic biomolecules and use these to enhance our understanding of evolutionary principles or structure function relationships of bio molecules. And so there is this this circular relationship here where at the heart. We have this interest in structure function and our ability to understand structure function is driven by evolutionary theory which it as you know is a descriptive or explanatory science and on the flip side we have this interest in biomedicine in biotechnology which is obviously an applied or a translational science. So we want to be able to take research themes over here and apply it to structure function analyses that help us develop evolutionary theories as well as take known standard evolutionary theories to understand structure function relationships that help us develop novel bio mom and kills. So at the heart of what we do are. Our sequence analyses biological sequence analyses and so here we have what my Ph D. advisor like to say is nothing more nothing less than a simple linear string of chemicals and. But our ability to extract information out of that linear string of chemicals is what's important for our laboratory and specifically we're trying to find some sort of information that tells us about the function of this protein or maybe where this protein came from from a historical or evolutionary perspective. And so you can imagine though that by itself. This really isn't very informative right. And it's not. And so you can put the sequences within a comparative and. Structure does this allow. To extract information that is useful. So what we would do then is we would collect what are called tamale guess sequences. OK these are sequences that share a common ancestor. And you collect as many sequences you can and it's not really. And so you collect enough sequences that you are able to extract some sort of useful information and. So for example in this one. Maybe someone already extracted something from there. But you can see you can highlight some sort of information within a linear string of chemicals so to speak. OK Not just because you can extract or highlight some sort of information doesn't necessarily tell you what its function is or where it came from. OK You just know that each one of those three species has a gene and that same gene believes that this is a true statement. But you don't really know why that gene maybe feels believes that way. So to speak right. So we want to do is we typically put these sequences within some sort of historical context and from a evolutionary perspective what this means is we try to analyze these sequences within a phylogenetic framework. And in the phylogeny then is allows us to analyze sequences within the evolutionary history based on extent organisms to understand their shared or common inheritance and so we can ask which extent species evolved from the same ancestor and how were ancestral traits modified in different lineages and so for those of you. Maybe that are coming from chemistry or from by a molecular engineering. This is a common way to represent relationships within biological sequences OK So this is the so-called phylogenetic tree where and this tree we would have extent. Org. Isms represented by these nodes here what are called terminal nodes or leaves or tips of the tree and calm. And these terminal or modern or extent sequences will share common ancestors at different points in the tree. So here we have these two common ancestor these two modern organisms share a common ancestor here. These three modern organisms share a common ancestor here and then there's the common ancestor for all five sequences or organisms on this tree and they're related by. Edges of the tree called branches and. So this really forms the heart of what we try to do we take biological sequences. We put them in a phylogenetic context and we try to extract information within that context. So just to give you. A brief example of how phylogenetics can be important. There is the classic case of an example where a where a doctor had injected his mistress. With with a needle that had HIV positive blood in it some time. And so this was from one nine hundred ninety four in Louisiana. In a woman claimed that her ex lover injected her with HIV blood in the records had shown that in fact the physician had drawn blood from an HIV patient that same day that the that the lover claimed that her. Her physician lover had injected her and so we want to be able to prove her demonstrate that the blood was from an HIV patient ended up in the woman. OK so how do we do that. So some background information about H I V E is we know that HIV has a high mutation rate which can be used to trace passive transmission or evolutionary past associated with the evolution of that with that with those gene sequences and so this is assumes then that two people who got the virus from. Two different people will have two very different HIV sequences sequences. And so we can do is we can try to build phylogeny S. OK we can build these biology trees based on HIV sequences to be able to attempt to extract something useful from some of these sequences and so what people did was they took multiple samples from patients the woman in controls and they built these phylogenetic trees and what they noticed is that there is this so-called nesting of the victim sequences within the patient sequences that the doctor had extracted HIV from indicated that the direct direction of transmission was in fact was from patient to victim and so this was the first time that phylogenetic analysis was used in a court case and care and so here is the tree. Right where we have this green cluster was the eight HIV patient that the doctor was treating and here we have the sequence from the from the X. lover and you see here this is an expanded analysis of the initial analysis as you see here that all of the. The victims or the X. lover sequences Claytor fall in our nest within the single patient's HIV sequences and. This. It could not be done with just these two it. You need directionality it demonstrates and it also demonstrates the probability of chance alone in two sequences being related. All right so. So that was just to demonstrate how important a phylogenetic analysis can be and how potentially maybe we can use a phylogenetic analysis so let me let me now talk about more specifically what we do with a phylogenetic and. This and how we can use that to extract particular information that's useful for our lab. So there's this concept in molecular evolution called Rate heterogeneity in cats and one form of rate heterogeneity is called hetero tacky and what this what this what. Or what rate heterogeneity allows for is an equal mutation rates among sequence positions within a within a collection of aligned sequences. Right. And specifically given that pattern you can have something called hetero tacky and what this model allows for is the mutation rate positions to vary in different branches of the evolutionary tree right. So that doesn't make sense. Probably less so let me explain exactly what this means so rate heterogeneity Here's an example of rate heterogeneity we have a single gene let's say and we have humble logs from four different species in comp. And what you see here is that the first two positions are slowly evolving OK they don't change at all throughout the evolutionary history of these four species but you notice for the third seat the third position. You see a moderate amount of mutation or evolutionary change and that's evident by here you have conserved hydrophobic property associated with these amino acids and so for the fourth position again it's conserved you don't see any change here in the fifth position is rapidly evolving. It doesn't seem like there's really any selective constraint on what amino acid is occupying that position and so if we were discovered magically represent that we would show this here where we have site one just has a single circle in it or a single ball to represent a low evolutionary rate. Same with position to position three however has a moderate rate represented by two balls so eight for a single ball and the last site is rapidly evolving so it is the most balls in it. So. The way that evolution. Is typically modeled up until. My Ph D. work was is that you assume this process was homo genius. And what that means is let's say Here again we have this collection of sequences OK So this is not a single sequence let's say this is the last common ancestor of life and this is represented represents a gene within that population of say tens of thousands of individuals and you're looking at the mutation rate at the individual sites within this gene in the population. So the first site is slowly evolving the second side is moderately evolving the third side is slowly evolving in the Forsyte is rapidly evolving. What we don't typically assumed is that this process stayed the same throughout time so as the last common ancestor of all of life diverge to give rise to bacteria and you carry out those individual mutation rates stayed the same throughout time. So the first site remains slowly evolving in both bacteria and you carry outs and the last eight remained rapidly evolving in both bacteria and you carriers. And so what my work showed was that well in fact this is not a homogeneous process rather this can change throughout time. And so you start with this example. Again here where we have the last common ancestor we have the same distribution of mutation rates but what happens now is as bacteria evolve from as they split from you carry out the first site goes from being slowly evolving to now being rapidly evolving in cat converse Lee the fourth site goes from being rapidly evolving to now being slowly evolving and so we've seen this shift today and rates and presumably that shift is somehow related to a functional constraint. That's acting on that gene and as it differs between bacteria and you carry. Gene presumably is that is doing something some small minor changes in its functionality or its properties between these two groups of organisms. So what happens though is that you. It's it can be difficult to detect this signal because what we typically do in the field of molecular evolution is we just identify a whole bunch of sequences we throw together we build multiple sequence alignments we build. Phylogenetic trees from those multiple sequence alignments and so since we're not since there's no bifurcation or dichotomy in the analysis of those sequences we can often miss these patterns. OK So if this pattern existed. Normally what we do is we just take a bunch of bacterial sequences and a bunch of you carry out its sequences thrown together into a multiple sequence alignment and if we were to look at the evolutionary rate of those sites we would not see that bacteria is rapidly evolving at that position. Whereas you carry oats is slowly evolving at that position rather that position would just look like it's moderately evolving. Encounter. So I spent my Ph D. career developing statistical tools to be able to identify this pattern here in cow and so let me just walk you through then. An example of the functional how that pattern is manifested at the protein level or the functional level cap so here's an example where we have a particular site. This site is slowly evolving in bacteria. OK So we have a conserved Argentine protein for the chemist biochemists and molecular biologists in the room. Whereas this same site is rapidly evolving and you carea. It's. So it suggests then that there's a strong selective constraint acting on the side in bacteria but that this constraint has been lifted or removed with. When you CARIO. It's. So when we compare a particular gene family in this case something called E.F.T. you between bacteria and you carry outs. Here's an example of the distribution of these types of what are called heter attack us. Sites sites that have a shift in their evolutionary rate between bacteria and you carriers so showing very. Are the sites that are rapidly evolving in bacteria but they're slowly evolving in you. CARIO. It's converse Lee the sites in red are rapidly evolving a new Kerio it's but slowly evolving in bacteria and. So we we distribute these sites onto a known. Three dimensional structure of the protein and what we're trying to do is be able to make functional inferences or functional explanations as to how these patterns can arise. So let me give you a brief background of exactly what this gene family does as I mentioned it's it's a gene family called E.F.T. you and it's a cheap protein involved in translation. Specifically what E.F.T. you does is it brings amino a solidity Arnie's to the ride the zone so E.F.T. you protein will bring all amino Eisa lated T. Arnie's to the Reiber zone when the correct answer. I code on code on base pairing takes place between the T.R. and A in the messenger R.N.A. that causes hydrolysis of the G.T.P. molecule that E.F.T. you in the spent nucleotide are then released from the ribosome in care then something called a nucleotide exchange factor in bacteria. This is represented by. E.F.T. Yes this nucleotide exchange factor kicks out the spent nucleotide in allows for a fresh activated nucleotide to rebind the to you. Molecule. And so. What we knew at the time is we had a structure for if to you. We did not have a structure for the homologs in you carry outs but we also knew where the nucleotide exchange factor bound onto the you. Camp. So we wanted to do is be able to make these explanations functional explanations as to what's going on with this distribution and so you can see here for example we have this collection large collection of red sites that are in this area here and this is where we knew we knew that the nucleotide exchange factor bound to you. OK. So these red sites the remember these red sites are rapidly evolving and you carry us but they're slowly evolving and bacteria. So we inferred then if they're slowly evolving in bacteria and we know that they are interacting with the nucleotide if changed after that the same sites are rapidly evolving as you carry out so that it suggests that the nucleotide exchange factor is not binding in the same way in bacteria as it is in you carea it's. And so we made this prediction we made additional predictions as well we predicted that this area here so these are Remember now these are green sites these green sites these are rapidly evolving in bacteria but slowly evolving and you carry out some K. we predicted that this area here would actually be an alpha helix. And you carry outs. We predicted that this is probably a binding domain and in you. CARIO it's because these these sites here form a hydrophobic conserved region and you carry oats. But these sites are rapidly evolving in bacteria which is what you would expect for something on the surface of a protein. So if something's on the surface of a protein and it's not really doing anything in terms of binding ancillary proteins then it's probably more rapidly evolving. But and you carry out these sites are slowly evolving. And so the surgeon off when our paper was and press. Somebody had determined the crystal structure for the you carry out a factor and our predictions turned out to be correct this region ended up being an alpha helix. And this region is where the nucleotide exchange factor binds in you. CARIO. Some cat so a completely different binding region for the nucleotide exchange factor even though these blue proteins are performing the same function within bacteria and you carry out they have a completely different way of doing that function. Who cares so. What hasn't. What's never been demonstrated up till this point though is nobody has ever experimentally validated this concept I'm so so this is what we are currently doing in the laboratory we are trying to take these sites mutate specific sites to be able to engineer these elongation factors from doing from binding there nucleotide exchange factors between the two different domains of life the bacteria and the you carry out some kind. So I don't I'm not going to show you any results because we don't have any results but I'm just going to set the stage really so that in the future of when we talk about this or someone from my lab presents us you'll at least have some background information as to where they're coming from and so here we have the nucleotide exchange factors between bacteria and you carry out these are now the known structures of these so in red. Here is the E F T U N. Bacteria and then you carry it's unfortunately they don't call it E.F. two you they call it something else. E E F one a K. and that's what shown inside and here and then we have the nucleotide exchange factors for bacteria shown in green and for you carry out shown in yellow. And so we want to do is we want to identify these header attack as sites that I talked about previously and we want to see if we can swap out how to attack a sites between bacteria and you carry out some care so there. Going to be sites that are going to be sites that are conserved in bacteria that allow the E F T you to interact with the nucleotide exchange factor in F.C.S. those same sites. Since they don't do anything over here. Those same sites are going to be rapidly evolving. So what we want to do is we want to take the site from bacteria and put it into you. CARIO. And we'll do this for a whole slew of sites and with the intentions then that we can get this bacterial E.F. to use this red protein to bind to this yellow protein or the nucleotide exchange factor from you. Carrots OK So we want binding partners to swap with the fewest or the minimal number of mutations possible where those mutations are driven or dictated or highlighted based on this evolutionary analysis. And so here's here we can align this just demonstrates that we can align structurally aligned E.-F. to U N E E F one a. And so these are the sites that we will be mutating in E.F.T. you in our attempt to to get them to bind to get E.F.T. you to bind the nucleotide exchange factor from you carry out some care. So these are all header attackers sites. These are sites that are rapidly evolving in bacteria but slowly evolving a new Kerio it's and so we are changing these sites in bacteria and with the intentions or hopes that they will allow then the E F two you to bind to the yellow protein. Converse they were doing the same thing in bacteria or the E F C You are in the one a and you curios were mutating that protein to be able to get it to the buy into the green bacterial nucleotide exchange factor. T S OK and here are the distribution. Of her attack US sites and in this example here. So this is really just setting the stage then for hopefully what our exciting experimental results that will you will be generating in our laboratory shortly. So let me shift gears then I'm going to try to shift gears a lot here today as I mentioned I have nine different research projects in my lab and I don't have time to talk about all those nine but it's not going to keep me from trying to talk about on this night. So another thing that we are interested in is being able to engineer or exploits the protein translation system in an attempt to develop therapeutics or novel proteins that contain unnatural amino acids and kind of so this is a. Research team that's of particular interest to the therapeutic community or the biomedical community because proteins that have unnatural amino acids have a lot of therapeutic value. So what we want to do is we want to manipulate specific interactions within the protein translation system and that are and those are interactions between transfer R.N.A.'s and elongation factors that bring those transfer our Nase to the ribosome. And so this historical view though of T.R. Ney's is that they are interchangeable components and passive adapters of protein translation. And there's this new view though championed in particular by human back at Northwestern that states that no rather tyrannize or active molecules adapted to meet specific kinetic requirements of protein translation. Specifically what that means is that the continent twenty T.R.N. a amino asal you know acid pairs have similar thermodynamic contributions. When they buy into E.F.T. you. And so when you miss a slate you put the wrong amino acid onto a tiara Nay it can have a broad spectrum of thermodynamic contributions to E.F.T. you binding and so what you love. Backs laboratory did was they essentially miss charged a whole bunch of T.R. Ney's with a whole bunch of different amino acids and then they looked at those those missed charge T R M A's in their ability to interact with you. And what they found was that there is this dichotomy between whether or not. E.F.T. you recognizes the amino acid or whether or not it recognizes the T.R.A.. And that this relationship is inverse So if you have a tiara ne. So here let's take the glutamate T R N N. If that we are in a interacts strongly with E.F.T. you then the amino acid group and that are in a the glutamate itself. That does not interact with the T. you molecule. Converse fully if you have an amino acid on the T R N A. That interacts strongly with the E F T U. So glutamine in this case that are in a does not interact strongly with the E F T U K. And what this allows then is that a particular thermodynamic compensation between the T.R.A. the amino acid in their interactions with you have to you right. So if you were to have if you were to take a one week amino acid in charge it onto a week. T R N A. Then presumably that Mish charge T R N A does not bind strongly enough to E.F.T. you in order for that charge tyranny to be delivered to the ribosome. So what the. That means then is if we were trying to make proteins that incorporated or contained unnatural amino acids as a first order what we'd want to do is we would want to buy an unnatural amino acids to these so-called strong T. Arnie's because presumably E.F.T. you can't bind strongly to an unnatural amino acid. So if we stick that unnatural amino acid onto a strong tear on a then the strong tarnation be able to deliver that to the ribosome. OK converse what we can do is bind unnatural of you know acids to weak tea Arnie's and simply hope that the E.F.T. you molecule combine that unnatural amino acid with similar thermodynamic affinity as the continents amino acid. The so you can imagine that that there's limited combinations of non-cognitive T.-R. name you know acid pairs. So a higher order rule it would be we want to enter convert a week. T R N A into a strong T.R. name. And this would presumably allow us then to stick anything onto any T R N E Right. And so what we want to do is we want to be able to explore if this observation that you have back noted is that there is this G You base pair present in the weakest as well as the strongest T.-R. names but they really haven't established any specific rule as to the distribution of that she you base pair. In the T.R. in a molecule. So what we want to do is exploit our evolutionary models to determine sequence structure function relationships of T.R. unease in their interactions with the two you. So specifically this G You base pair as I mentioned what it does is it. It generates or presents an increase the lecture a negative. Potential to anything that is juxtaposed on the E.F.T. you molecule to interact with that increased a lecture and negative surface. And so what that could mean then is if you. Had a positively charged of you know acid on you have to you. It could interact more strongly with this bass player than say it would with a G. C. base pair in can and so we did was we we analyzed T R N A sequences and we have this again we have here's this phylogenetic tree or this this clay gram that represents relationships of the sequences and this. Clustering is based on the distribution of the G. You base pair and what we noticed is that these so-called strong binding T.R.N. A's have a G U in a specific place on the structure of the T. R.N.A. whereas the weak binders have achieved in a different area of the T.R. and then. So specifically you have the G. when the G U is around here in the except or stem. Those are the strong binding T.-R. Ney's whereas the weak binding T.-R. in A's are have a G U in the TS down but size so what we want to do then is mutate individual sites based on our evolutionary analysis so here's an example that we are working on we have a T.R. in a veiling OK Valium is one of these we are nice. So what we want to do is convert this week. T.N.A. into a strong binding T.R. next comp. So here is the G. you. So the G. us in the in the T. stem it's flipped this orientation is flipped from the previous line. We're now looking at the mirror image of the Tiernay So here we have this to you and what we're doing is we're mutating not only the G U but we're also introducing a G U into the except or stem because that's what we noticed to use happen to be for the strong T.R. Ney's So in this example we've taken the veiling T.R. in a the weak binding T.R. any where introducing a G U in the accept or stem appear. As well is right here and we're also have examples where we are getting rid of the G U in the T. stem. And so the way that we're going to measure this is where using something called in the your system which is a reconstituted in vitro translation system where you are isolated all of the individual components of protein translation except for the ride you isolate that from an organism so recombinant Lee purify express all of the individual protein components of the translation system in this case we're using eco like components and this is a collaboration we have with Matt Hartman where we can then manipulate the system OK so instead of putting in say wild type T R Nase we can put in our engineer tyrannize and look at the amount of protein that's being made in this system. OK So again here we have we're working with the valium T R N A which is a week. And we're not yet working with the natural amino acids what we're going to do to test our ideas stick different. Natural amino acids on to this veiling T.R.A. So we're going to stick a week. Amino acid and a so-called strong amino acid on to this week. T R N A So this combination of the valium T.R.A. with the glutamate weak amino acid you want you want suspect that this would participate a lot in protein translation converse Les you would expect this pair to participate. A lot in. Protein translation because now here you have a strong amino acid with a weak T.R. name. And so here is an example. This was the the this is our first ass assays that we've done so far and what we have here are two groups we have the example here where we take the glutamate and we. Stuck it onto the. Veiling T. R.N.A. using a rubber arrives on time and then we take the glutamine amino acid and stick it on the veiling T R N N And so you can. So although these are preliminary results you can identify some general patterns that are occurring here OK So when we take the wild type of alien T. R.N.A. and stick a week binding glutamate amino acid onto their this doesn't participate a lot and full length protein products into. Converse they when you start making these mutations into the veiling T.R. in a OK so here whether you're making it in the accept or stem the T. stem or in both what you noticed is that there is this increase in this ability of the mutated C.N.A. to participate in in protein translation so if you just get rid of the G. you in the except or sorry in the T. stem what happens is that molecule can then participate more strongly in protein translation in the combination of getting rid of the G U in the teach them as well as adding the G U in the accept or stem allows for this charge T R N A to participate really strongly in protein translation into converse the way you would expect in this case where you have a strong amino acid in a week. T R N A is that it really shouldn't matter what the mutations are in that thing. And besides this example here we know we have some air. You see in general that you have the strong amount of of participation regardless of what the mutation is and. So what we're doing is we're doing a lot more examples where we put different amino acids on to different to different mutated or engineered T.-R. names alternately with the intentions that we can get these things these engineer tyrannize we can stick a natural amino acids onto them and then they will be able to make proteins that contain unnatural I mean. Yes it's for us. So let me shift gears again here and now talk about this concept of ancestral sequence reconstruction that we work on in the laboratory and how we are exploiting this research team to allow us to better understand evolutionary principles a lot and also allow us to develop novel bio molecules. So here is a very brief history of Earth as maybe some of you know Earth formed about four point six billion years ago and for roughly the first seven hundred million years of our rigs. If the planets existed. We were being bombarded heavily by plan a test. That would have certainly vaporize any liquid water present on the planet and so. Shortly after this heavy bombardment and we know that the earth's crust solidifies the oceans form and then some time after this three point nine billion years ago. It's controversial but at some point certainly after that you have this evidence for bio signatures of life and can a lot of these are contentious that people argue over them probably one of the least arguable examples of of a a signature of life left in the rock record is this example here where we have a roughly three and a half billion year old rock from Western Australia and you can see these patterns in the rock here and it's believed that these patterns are created by organisms like this that currently live on our planet in Australia in the Caribbean. These are for those of you that don't know stream Adelitas they are essentially very large biofilms where hundreds of thousands of different species are living together in these clumps Kyra. And so what we want to know is we were are interested in what kind of environment of these ancient organisms. Live in. And so the field of Molecular Evolution offers a clue in terms of the structure of the so-called Tree of Life where here we have. The Tree of Life with the three domains of life bacteria Arky and you carry out and this is roughly believed to be the overall distribution of life on that tree and shown in red. Then the red branches are these are what are called hyperthermia Philip organisms OK. Hyperthermia files are defined as organisms that grow at or above eighty degrees Celsius OK so these things love living in hot environments converse really that would be compared to organisms called thermal files which grow between forty five and seventy five degrees Celsius and then you have maze of files and Saker files which grow at even lower temperatures. So what based on the distribution of these hyperthermia files on the tree here. It's believed that the last common ancestor of life for the last common ancestor of bacteria also lived at a hyperthermia Philip environments or even a very very hot environment such as in our days you can find these things called black smokers or hydrothermal vents on the ocean floor. So typically this notion of understanding ancient environments is conducted research by geologists and. So we want to do is we want to be able to pretend like we are geologists and go back in time where instead of looking at a rock. We want to look at sequences biological sequences and try to extract some information from those biological sequences that give us a clue as to an ancient environment and. So the analogy that we typically talk about in our lab is that of what a historical linguist does a historical. Linguists will look at the relationships and spellings of a modern word. So here we have the modern word for snow and based on this distribution of the modern word you can estimate what an ancestral spelling of the word snow is or in incest or pronounce CA shin of the word snow. So for in this case here we can go back five thousand years to the proto indo Europeans that gave rise to these modern languages but in this case the reconstruction says something about the proto into Europeans right in their environment. Specifically they lived where it snowed. So we want to be able to do the same thing but with biological sequences instead a man made words and so we're going to have this transition essential to going from the human alphabet to nature's alphabet we use that same thing that a Sturrock a linguist does the same type of analysis but use biological sequences place and within a historic goal context or an evolutionary tree and then estimate what the ancestral spelling of that gene would be. So in this case then we would just replace languages and words with a particular gene and we would estimate what that gene sequence would be a particular nodes in that phylogeny So we are reconstructing the ancient sequence associated with this gene in Conn in our case were interested in something that goes back to about three built more than three billion years ago with the dawn of bacteria. So what we do what we have done is we have this phylogenetic analysis and this phylogenetic analysis allows us to infer the ancient sequences at the individual nodes here. OK this is all computational and then we actually build those sequences in the laboratory we recommit Lee express those genes make the ancient protein in laboratory and measure some sort of function associated with that ancient gene. OK. Again I'm still working with this elongating factor to Eugene family it's turned out to be wonderful for our lab and what's nice about this gene is that it essentially serves as a molecular thermometer. So the gene the protein creates a thermo stable protein when that protein is isolated from thermal fillets organisms. But whereas that protein is not thermo stable when it's isolated from Misa feel like organisms converse Lee that you have to you from thermal files is not optimally functional it Misa Phillip temperatures in camp and so there's means there is this linear relationship between optimal binding temperature in optimal growth temperature of the host organism that has that protein can. And so for example we have a visiting scientist in my lab Michael Graham who's the world's expert in structural bioinformatics has already shown demonstrated this relationship with other proteins and can. There's this near linear correlation coefficient between the optimal temperature of the protein and the host temperature and of the organism. OK so here we have this thermometer. So what we want to do is we want to be able to. Look at the specific thermal stability profile of the proteins whether they're modern or engine to give a clue as to what type of environment those proteins come from. So here's an example this just to show you this relationship in modern organisms. Here we have the E. coli pro-team we are using circular dye Crowism to look at the stability of that protein and so as those of you that work with the color you know that it grows up to million about forty degrees Celsius. And so this temp this and the melting temperature with the E.F. two you protein is right around the optimal growth temperature of its host organism. OK The same thing with the protein from thermos aquatic. Growth temperature of thermos is around seventy five degrees the melting temperature of the protein from that organism is around seventy seven degrees. So we did was we generated a phylogeny OK here's this trees again and we reconstruct computationally reconstruct ancient sequences at nodes of this tree and then we build these sequences in the laboratory recombinant Lee express those genes and measure the thermos to build the of those ancient proteins and what you notice here this general trend is that as you go further back into time whether you are starting out with these modern organisms or whether you're starting with these modern organisms the further back in time you go the more thermal stable the ancient protein is. And so that allowed us to correlate this temperature trend to known geological evidence for the ancient ocean temperatures. So here we have an estimation of the ancient ocean temperature based on the fractionation of oxygen isotopes in the in the rock record so if we go back three and a half billion years ago the ancient ocean is predicted to be very hot. This is degrees Celsius here on the Y. axis and cat so eighty to eighty degrees Celsius. It's extremely hot for us at least. OK And as you approach modern day the present time you have this near linear decrease in the temperature of the ancient ocean temperature of the ancient ocean. OK Here is another geological based study looking at a different isotope and you see this same general trend. And so we were able to do is we were able to put time points associated with our ancient proteins where here we have our tree again and we have different nodes that we reconstructed on this tree and on the axis down here is we have time. So the blue dots represent all the different nodes that we reconstructed in ancient enough to. You and then we have the time associated with those divergence dates those nodes OK So that allows us to we have temperature and now we have time now so we can plot our results onto this same type of graph and what we noticed then is the same general trend where this very ancient life lived in a hot type of environment or at least the ancient life that gave rise to modern life lived in a hot environment and that that. Those organisms progressively adapted to their changing environments as would. To be expected under the paradigm of Darwinian evolution. OK So let me then just spend the last couple minutes here talking about how we are extending this work. OK So one thing that we are trying to do is be able to use this information to understand how thermos debility evolves and proteins. Specifically we want to know if we can exploit this information to build models that attempt to engineer there must a billet E. in different protein families and this is in a lot of ways. This would be kind of the Holy Grail for biotechnology in industrial processes because often these processes require could tell us this at elevated temperatures for enhanced product substrate solubility viscosity and could tell us this is well as to prohibit biological contamination so if we are able to develop an algorithm or model that allows any protein to become more thermo stable than it currently is then that would hold terrific value for those communities and so we are doing right now is we are looking at the structures of these ancient to you. Proteins in collaboration with Eric or Atlanta crystallographer at Emory and work with Michael Graham Ian in my laboratory where they are crystallize So here is example of aligned E.F.T. ancient E.F.T. use structures that we've crystallized in the term in the three dimensional structures. And so what we want to do is we want to be able to use this information to help us develop models where what's interesting in this case. So we have let's say an ancient protein here an ancient protein here. We know the amino acid changes that occur between these two nodes on the tree. We want to be able to predict which of those changes are responsible for that change in thermal stability associated with those ancient proteins ultimately being able to take multiple passes along this phylogeny and look at these structures and make predictions and then actually test these predictions in the laboratory and to make a database that could be exploited by others with. Different interests. Another thing that we are trying to do is we're trying to replay this so-called tape of life. This is based on this notion famous evolutionist called Stephen named Stephen Jay Gould at Harvard who made this comment that any replay of the tape of life would lead evolution down a pathway radically different than different than the road actually didn't really say that the rather than the road actually taken if we could replay the tape of life. OK so what we're doing is we're taking these ancient sequences these ancient genes engineered to you. And we are sticking it into a modern organism or we've gotten rid of the indulge in this protein can. So for of all the thousands of genes in any coal lie. We can knock out one of those genes stick in the ancient one and watch it evolve in the organism. And what we know is that there's going to be this strong selective pressure for this ancient E.F.T. you molecule to accumulate mutations to change it's there must ability profile when it's evolving in interacting with in the whole Y. organism. In addition we're trying to address this criticism where how do you. You know you've made the right. Ancestral protein right we don't actually know unless we can talk to God unless we can time travel. We don't know what the true ancient sequences are. And so we want to do is we want to be able to benchmark and special sequence reconstruction and so this is what Ryan Randall in my laboratory is doing where we are taking a red fluorescent protein. And we are randomly mutating it and we are selecting for a variance that have particular phenotypes and so what you can potentially imagine then is we can take this sequence we can evolve it in the lab we can evolve it in a way that is. That is similar to how evolution works so we can evolve it along this tree like structure in Qana so we can mutate the sequence we can look for a variance say for example we can look for variants that maintain the same phenotype we can look for variants that have new phenotypes and at the end of the day we're going to have a whole bunch of sequences. We're going to be able to take those sequences plug them back into our computer algorithms that predict engine sequences and we're going to then rebuild the ancient sequence and see if we can recapitulate this green phenotype this red phenotype this red phenotype in this red phenotype So this allows us to benchmark and several sequence reconstruction algorithms and. Also minutes and Sestriere sequence reconstruction would be not just a gene not just a pathway. But a whole organism in crime and this is what simming Joe is doing in my laboratory where she's trying to computationally reconstruct a complete bacterial genome with the intentions of exploiting recent technological advances at the Venter Institute where they have not only synthesize a complete bacterial genome. But they've also transplanted one bacterial genome into a different host species and then that host is able to divide. And one of those daughters has the synthetic genome and so this technology that is is essentially right for us to be able to take advantage of to potentially create an ancient organism. OK So she's doing this with Michael plasma. Another research project in my lab this is work being done by Megan Cole and what it relates to is the origins of life and so I like to for those of you that don't know Nick Hyde is is Georgia Tech's resident expert in origins of life and what I like to tease next because there's this notion of the are in a world that predates the modern world and that I like to say that the R.N.A. world is so-called hollow at the core because people are trying tried to use our in a as the single bio molecule for life and. And what that means then is that there's this potential paradox because life as we know it today is a two by a polymer system where you have information and could tell us this being done by two different molecules here you have D.N.A. Here you have protein and so there is this paradox How do you have just a single molecule doing acting as both the information content of life as well as they could tell us of life. Now of course we know that this happens in real life today with the form in the form of Privacy Times but in there are in a world. We have this we have this potential situation where maybe there never really was just are in a around without protein. So we know is that based on interstellar clouds meteorites the Stanley Miller experiments. There are lots of amino acids around. And what we also know based on random are in a library says that you can identify ribozymes that catalyzed peptide on formation. You can identify random are in a molecules that can catalyze amino a salacious of T.R. ne. So really what this does is it essentially it can argue for a scenario where you have these statistical random peptides just lying around in early life and so you can ask potentially Well is there a scenario where a particular type of pool of random peptides would have contributed to in some advantageous way to early life and so that's what Megan Cole is working on in my laboratory where she's trying to understand this transition from an R.N.A. world to a to a ride both nuclear protein world in case you've taken two approaches where hopefully those two protests will meet up in the middle with what we're doing with the Georgia Tech NASA Astrobiology Research Center. And this well we have Josh Dern who is working on trying to resurrect a whole ancient Ribfest zone where the last common ancestor Schreiber's zome and this would be analogous then to say what NASA is interested in is this planetary exploration. Well what we're doing in our centers the Reivers So Max floor ation we're taking lots of different approaches really to understand how the ribosome has evolved some calm so. Lastly Dan is. So as I've shown here today we have a lot of different research teams in my laboratory and my students like to kiddingly argue as to which research is the most exciting but the one research team that has a particular interest to me is a project that James crasser is working on because my Ph D.. I was trained in a bio medical fields and K. but I Applied Molecular Evolution to the bio medical field. So the ultimate would be to somehow generate a therapeutic potentially and buy for biomedicine but that is based on evolutionary analysis or evolutionary work and so that's what James is currently doing unfortunately. We don't have patent protection on this yet so I can't publicly disclose what we're doing since we're being video recorded here but I would encourage you to. To corner James at some point and ask him what he is doing and so with that kind of anti-climactic. Explanation they're going to end in thank you for your time this morning thank you thank you. But with zooms like a lot of what you're doing. As. Well. I mean. Beyond just making. Proteins that have unnatural amino acids we haven't given any consideration yet but you know that's why I'm at Georgia Tech's there's collaboration's all over the place and so hopefully that's a sample that's right for the picking. Steve. Yeah like. Most of what they are and yet there are differences some kind clearly but so what we've done though is that we expect. So we have this you know the spear the surviving term sphere that rolls around and tries to if you're within that sphere and you are a head or attack a site then you will potentially contribute to this change and in binding partners but we can't you know the worst case scenario you're right as we introduce these changes in the second the tertiary structures are so different that those sites are not going to be able to interact with the complimentary nucleotide to reverse. Yeah I haven't I don't now I don't know what it takes to to shift to the tertiary structure. But we are making different combinations of sites. So maybe we will be able to see this shift occur over time since we're collaborating with the crystallographer we would. Crystallize these things. Yes. You're right. The short answer to that is know who people have been trying to identify some common property that you can exploit to convert a non-thermal stable protein into a thermo stable protein. They've been trying for years. No one has really been successful. You know we can we can hack at it a lot until we can get a protein to be thermo stable but it ends up not really following the same path that we set out to do and so but that's what we're trying to create To be honest I don't think will be all that successful but I think we will make some contribution towards it. That allows us to take a step in that direction. Roger. There was a sense. So you're asking about sub header. Taki. You care. Right. So you. So we've never looked at hetero Taki within a say domain of life. We've never looked for had a techie with in bacteria presumably it's there because there are major bifurcate. It's within the bacterial domain. But no single the way that the algorithms work. No single sequence is going to mess up our analysis. So if we had a completely conserved aspire to it and all of bacteria except thermo toga had a trip to feign that's not really going to affect our analysis because it takes branch legs into consideration. Yes Sujan last question. Right. So not not with this gene family. There's a few examples. So even in fact there are different elongation factors. And this gene this family is not one of the ones that experienced a lot of gene duplication and it's also the reason that we're able to exploit this family so much that it's also one of these families that doesn't appear to be horizontally transferred at all because that would also really screw up the analyses. But potentially you could arm. If you did have duplications events as you know you're going to have you could have if there is say sub functional the zation between those duplicate and that would certainly obscure. The signal of functional divergence. But in fact you in that case you would be able to look at para logs with this functional divergence technique in be able to identify the sites potentially that are responsible for the sub functional is ation. Right. So with that. Why don't you get on with your days and thank you for coming out this morning.