 My name is Eric Jarvis. I'm a professor at the Rockefeller University in New York City, full professor there, and I run the lab of neurogenetics of language. What brought you to science? So what brought me initially to science is I was just fascinated by how nature works. I wanted to understand, you know, the mechanisms of life itself as well as the origins of life and civilization. Can you tell us specifically how dance and the arts have influenced your scientific work? So a lot of my family, particularly on my mother's side, but my father's side as well, they were into the arts, being musicians, some of them dancers, some of them drawing artists. And I was on that path to become an artist like them. And I wasn't that great at singing as some of my relatives were. I ended up being better at dance. I once had a high school performing arts in New York City as a, you know, during my high school or like teenage years and Alvin Ailey and Joffrey Ballet. But at some point I fell in love with science while I was a high school student and I was trying to make a decision at the end of high school as to what do I want to do with my life. And my mother always said, do something that has a positive impact on society. And I felt I can do that better as a scientist than as a dancer. And I made that switch just automatically overnight that I wanted to do something that has a good impact on society and that I loved doing, which was science. And I discovered over the years though that being trained to be a dancer trained me to become a scientist because they both require a lot of creativity, a lot of hard work, a lot of discipline. So I encourage people who are into the arts that you're actually being trained to become a scientist as well. And if you want to switch, it works. What are some of the benefits of being an artist when it comes to science? I wouldn't say there's no distinction between the sciences and the arts, but I would say they are in a continuum. And that the kind of things you need to do, the kind of personality you need to have to become an artist. I'm going to say overlaps with that of becoming a scientist. To be a successful artist, you're trained to persevere. You're trained to try to overcome obstacles. You're trained to accept getting rejected in a lot of auditions before you have the success. So I didn't see that was any different in science. You got to try and experiment 10 times to get it right once. You have to submit grant proposals, papers for publication and get rejections, and then pick up your feet and do it again and do it again in a different way until you get it right. Can you talk a bit more about your mentors and the importance of mentorship in science? Ernest Everett Juss was a scientist back in the 1930s in the US and also in Europe before I was born. And he was an African-American and very good scientist but had difficulty as you can imagine back in that time, even today, I would imagine, getting a job as a scientist. And I was a graduate student at Rockefeller at the time I first heard about him from another faculty there who was also African-American but not a tenure track professor. And he gave me Ernest Everett Juss's biography to read and it really resonated with me. So here is a person mentoring from the dead, so to speak. It resonated with me because here I am at the Rockefeller University where he had a hard time landing a position and now I'm a student there, one of the very few students of color. And I could identify, even in the 1990s, what he was experiencing in the 1930s. And I used that ability to identify with him almost as a self-mentoring from the past to decide how I handle my experience at Rockefeller. Mentoring broadly though, I've discovered that there are certain peoples, you don't have to have somebody of your ethnicity or of your gender to be a mentor. So when I was at Hunter College and I first started doing research in the laboratory of Rifka Rudner there, was a Jewish woman, Caucasian, and really took me under her wing and taught me not only how to do experiments but how to think like a scientist and also just gave me feedback on life experience and how to balance between work and life. And she eventually would say she's my Jewish science mom. And so different ethnicity, different race, she was one of my most valuable mentors and helped me get through things of even of issues of diversity. How was your experience at Hunter College different from Rockefeller University? It was a sink or swim feeling at Rockefeller and the transition from Hunter College. And my advice to folks is that treat it like you're going from one culture to another culture. All right? Even if there are a few blocks away in the same city, there could be different cultural experiences. And you have to treat it as if, okay, you're in a foreign land, you're not familiar with this. You might feel isolated. You might feel like people are speaking a different language. So take the effort to reach out because people aren't always going to reach out to you. Reach out and learn from this new culture and how to either integrate into it or get them to integrate into yours. Do you think that a sink or swim mentality fosters better science? I don't think the sink or swim environment is better. And it could matter to a certain degree, yes, the types of personalities going into that environment. But there's so many talented people out there that I've seen come from a more nurturing environment that are doing good science that could contribute to society, go into a sink or swim environment and fail and don't do as well and don't come out of it. So because of that, I wouldn't say it's better. I wouldn't say more competitive, you know, me, me kind of environment. And some people will okay in that environment, but not everybody. How did you come to your specific field in biology? What questions were you trying to answer? Yeah, so I started out in science and bacterial molecular genetics because I was just interested and fascinated by the area of molecular biology, playing around with genes, playing around with molecules. So once I got over that fascination as an undergraduate student and I got into graduate school, I became fascinated with two questions. One was, you know, beyond biology, the origins of the universe, how did life begin as well? And the other was how did the brain function? And I chose how the brain functioned as my big question because it controlled dance. And it's something I can touch and it's something I can relate to. And then I narrowed it down further by saying, well, I'm not sure I can really, there's nobody I know studying the neurobiology of dancing, right? And I want to choose an interesting question in the music world. I guess I can say, you know, in the music world, you learn to, or you're taught to learn how to play on a hard instrument first. And if you then master a difficult instrument, then a other instrument, a simple instrument becomes easier and you sound great. So I thought, okay, what's a hard question in neuroscience? And I thought, that's consciousness. So with consciousness, I'm going to stay away from now. So I'll choose the next hardest question and that's language. And that's how I started studying the neurobiology of vocal learning because language depends upon a highly specialized trait, the ability to imitate sounds and very rare in the animal kingdom, us humans, parrots, sawn birds, a few others, dolphins. And so that's where it got me here today. Why did you become interested in understanding the characteristics of vocal learning in animals, especially birds? So this idea of why is vocal learning so limited in a few species, the reason why that's the case, right? It came to me while I was sitting in a garden at Duke University, Duke Gardens, and I'm reading a paper and I hear this song sparrow singing from the top of a tree. It's song like, and it keeps singing this over and over again. When he first sings it, he stands at the top of the tree and is up there and look at him. And I stare at him a little bit and, okay, I get back to my paper. And then 10 minutes later, he switches the song, right? And then I look up again. He's getting my attention. And every five to 10 minutes, he keeps switching a song and I'm not a bird, but I'm paying attention to him. And I think, how stupid, right? He keeps getting my attention and the attention of all the other animals around him by stopping my auditory pathway from habituating to his sounds. What if I'm a predator, right? And I thought, wait a minute. Maybe the vocal learners have overcome the predation risk and are able then to evolve ability to get the attention of the opposite sex in this case, but overcome predation. And it turns out that the top predators tend to be vocal learners like humans, or let's say the top of the food chain, humans, dolphins, elephants. And then it turns out that songbirds and parrots, the vocal learners we discovered later, evolved from apex predators amongst birds that are related to eagles and hawks. So we think predation is selecting against vocal learning. Yes, that's right. The pull is sexual selection and complex social communication is selecting for vocal learning in spoken language and predators are selecting against it. And this balance then allows the species to become more complex vocal learners and the more simple vocal learners. How do you deal with controversy and complex questions in science? Some people think I make a big hypothesis just to spark controversy. That is definitely not the case. I come up with these so-called crazy hypotheses if someone called in that because they make sense to me. And sometimes the other hypotheses don't make sense to me. They sound crazy to me. So I really go by what makes sense, what's logical given the data, given the evidence. Now of course, this is conjecture when I'm hearing a bird sing from the top of a tree and so forth. So I go in the literature and I start to search, okay, who is an apex predator? All right, who's a vocal learner that's not a predator, not at the top of the food chain like bats, right? They're vocal learners. Oh, well, they're singing in the ultrasonic range where no one else can hear them, but their predators can't hear them. So it starts to make sense. I like to go upon the evidence. Now in terms of big question or big picture type of science, I like to push my science in the direction of addressing the big picture is because otherwise I get bored. And that's simply a matter, I mean, that is the primary reason why I try to go for the big questions. You've been quoted as saying, science by any means necessary. What do you mean by that? That statement I made, science by any means necessary comes from my upbringing and my parents' upbringing in the civil rights era. So this is a cultural influence and I grew up hearing my family argue, do you support Martin Luther King or do you support Malcolm X? And I mean, Martin Luther King was a person who we were taught and says, bring everybody together, love everybody, treat everybody as your brothers and sisters, whereas Malcolm X's approach was you can be segregated, but try to achieve by any means necessary. Now that can mean violent means and I don't mean this in science. What I mean is bring all techniques together and do what is necessary, do not be afraid to learn a new technique, to address an interesting and challenging question by any means necessary, learn it or collaborate with people. And those are the civil rights leaders who have influenced my science in this way. Do you think the funding structures of science incentivize that kind of approach? No, the structures of science do not capitalize on high-risk, high-impact. High-impact they want, but high-risk no. Or even somebody who doesn't know bioinformatics, who's, let's say, a wet lab scientist just taking the leap and trying to learn bioinformatics or vice versa, bioinformatician trying to do wet lab research. So what was striking to me is this is the way that I was taught growing up, dream big, like Martin Luther King, to try to change society. To me, this was normal. So to go through the grant review process and find that it's not normal was kind of shocking. But now I know it's not normal, but I still do it anyway. Because I think it needs people to change the system. It needs people not to accept the normal way. Are there some rules you follow when taking this approach in your research? Yes, I do have some rules that I've learned over the past 20 years now of being a scientist. And that is go for the big questions, but make sure you have incremental steps that can lead you there, such that if you don't reach the final goal, right, you can see it, but you don't reach it. You know that you can do some steps along the way, some milestones that can make an impact and then let somebody else take it from there. That's one rule for my students, for my postdocs, everybody that works with me in my lab. And the other rule is sometimes what's so simple and staring at you that other people consider, you know, and, you know, let's say. The word I'm thinking about is, doesn't make sense at all to them, right? If it makes sense to you and maybe to two or three percent of the people and not to the 95 other percent. If you are convinced by the evidence, sometimes that 95 percent will eventually be convinced five to ten years later. It's OK. It's OK to go at it alone for a little while, as long as that evidence is convincing to you. You know, I know we're also talking about genomics. I give you a genomics example is when I first started getting involved in producing genomes, I was convinced, based upon the evidence, that long reads is going to generate the highest quality assembly as possible without short reads. I had such a difficult time convincing the genome experts, and I'm not a genome expert. I was, I'm a neuroscientist. Well, maybe now I'm becoming one, but, you know, I started as a neuroscientist and it didn't make sense to me looking at the evidence. It's amazing how so many people, right, can miss something so obvious if they're entrenched in the field for a long time. Where somebody from the outside who's just coming in for the first time sees, well, this other alternative solution is the most obvious. Why is anybody seeing that? Because it's so easy to get blinded to the status quo. So my point here, in referencing your question, is be careful about you, me, other scientists becoming part of the status quo. Always challenge your assumptions. Do you think there's a risk that the field of genomics will become static as it continues to become more established? I think in genomics there's a lot of continued disruption that's going on that we're not, we're not at the point of total entrenchment, I should say. There has been entrenchment, but it's, there's enough disruption going on that I've seen the field of genomics change much faster than I've seen some other fields change. How did the completion of the Human Genome Project influence your research at that time? That's right. So when I first got into trying to study the mechanisms of vocal learning in humans and other species in songbirds, I was cloning genes one at a time from songbirds and eventually we wanted to see what their human homologues were like to see if humans was similar to songbirds in different from chimpanzee who doesn't imitate or different from chickens who doesn't imitate. So already from the beginning, even before we had whole genomes, my work was comparative. It required comparing species that have specialized traits with those who did not. I couldn't stay focused on one species. So I was very happy the human genome came out. And when the human genome came out, I was not a genome biologist. I was assuming, okay, they say draft, but I'm thinking it's mostly complete. Everything's there. It's maybe some difficult regions aren't there in the genome. But my language genes are going to be there. My vocal learning genes are going to be there. And we just needed to get the songbird sequence. So I got involved with a bunch of people involved in avian genomics. And the chicken genome was the first bird genome to come out. And then the zebra finch, a songbird to come out after that in which I was involved in. And the chicken and zebra finch folks were sort of marching arm to arm together. The chicken came out first though. But once that happened and we started looking at the genes, we discovered, wait a minute, these genes, genomes have problems to them, including the human genome. And then that's when we started getting involved in cloning individual genes over again and trying to bring in more species. And as we brought in more species, we couldn't spend a billion dollars or even 10 million dollars, which we spent that time on the chicken and the zebra finch. On all these different species, we needed a new paradigm shift. And that's when next generation sequencing came along. How did the advent of next generation sequencing change how you approached your work? When I say we in the neuroscience vocal learning community, in the language community, we have been able to identify the phenotype of vocal learning across multiple species and see who has it, who doesn't have it. Even it's a long and continuum. And so now we have a set of species that we need the genomes for. And I remember when the American Recovery Act in 2008 to 2009 came out that was helping scientists do their research without having to lose their jobs from the U.S. government. I put in a proposal. I want the genomes of hummingbirds, songbirds, parrots and their close relatives done so I can figure out the neurobiology of language. Just give me money for 10 species and I'll figure it out with this next generation sequencing technology is not going to cost 10 million dollars. It's going to cost, let's say 200 to $300,000 per species at that time. That was a paradigm shift. I didn't get the money, but we managed to convince other people to help give us the money through crowdfunding among scientists to address such a question. And I never believed in my lifetime I would have the genomes of all the vocal learning species I need and their close relatives before I die. And I had it in a matter of a few years as a result. From your perspective, was the pace of sequencing technology development really that significant? Yes. The pace was that profound because the zebra finch took, I was also involved in the mouse genome project. It took like four years. The zebra finch took five to six years, millions of dollars. And whole big giant consortiums of people and investment from multiple countries and so forth just for one species. So imagining that in my lifetime was hard. But once the next generation sequencing technology came along, linking back to one of your prior questions and made me think what's now semi-cliché is that we kind of overestimate what we can achieve in two years, but underestimate what we can achieve in 10 years. And this is what next generation sequencing made me realize it accelerated our ability to do all these genomes. And that's what then made me say when I started, you know, becoming more involved in genomics, let's plan on doing all bird genomes on the planet. Let's plan on doing all vertebrate genomes because it may happen in our lifetimes. Are you concerned there is a gap between technology development and the questions genomics researchers are asking? There's a gap. So one thing I've discovered about the genome world that's different from other areas of biology is that the genome community began with you only get one chance. You get one try. Whereas in my regular wet lab research in neuroscience, you have to replicate your experiments multiple times to be confident of your results. So the human genome somewhere between let's say 2.3 billion was I think is a real number that it costs. How are you going to replicate that multiple times to make sure you got the right genome assembly? And so there's this gap. It causes a gap. And that is because you do it once there could be a lot. There are lots of errors in your genome, in your assembled genome, not the real genome, but in the representation of it. And those errors cause lots of problems in analysis, in what students are working on. I mean just serious problems that people don't realize. And now with genomes becoming cheaper. You can replicate multiple times. You can experiment with different assembly tools on the same individual. But the genome community is still in the past. You only get one chance. So even with the cheaper next generation technology, they just tried once and then they move on. I'm not used to that. So what I believe I'm helping to bring into the genome community is replicate your experiments. You try to new genome assembly approach. You try to new technology. Do it multiple times, even if it costs five, ten thousand dollars each experiment. Do it multiple times because it's going to close that gap between low quality data and the people working with that data to produce new scientific discoveries. And one way to understand this is your results are only as good as the data going into it. And the genome community has to get used to closing that gap by making the data as good as quality as needed for the science. The questions that are needed to be addressed. Do you view genomics as a descriptive science rather than a hypothesis driven science? Is that distinction valid? So I would say I agree that a lot of genomics is descriptive. But I'm also going to say that I think hypothesis driven type of research is overrated and descriptive research is underrated. Further, descriptions help generate hypotheses. And second, I mean, and third is that I see myself and others bringing hypothesis driven thinking into the genome community, not even only for the basic science discoveries of how mother nature works. But how do you do a genome assembly? Someone tells me, well, if you combine this linked read 10 X genomics approach with this bio nano approach, you should get a higher quality assembly than if you combine it with high C or vice versa. And I will tell them, well, that's a hypothesis. Let's test it. Let's let's do all these different technology sequencing paradigms on your samples and test the hypothesis that this is going to generate. A better assembly for plan A than for plan B. And I don't think the genome community thinks well enough in that matter, even for the technology development. The hypothesis driven technology development is what I'm trying to speak of. What excites you most about the potential of new genome sequencing technologies and what questions do you hope they will help you answer? So I think some of the most far reaching questions are going to come from duplicated sequences or repetitive type of sequences. Those have been the hardest to assemble. And as I mentioned earlier in my presentation, that haplotypes are also a duplication. And one haplotype, the maternal paternal chromosomes are basically just one giant repeat that's hard to assemble. And mother and father chromosomes can be quite different. So we're going to see variation in haplotypes not only within an individual animal or individual human, but across the entire species. That's going to give us interesting biology about different phenotypes according to those different haplotypes. An example close to home for me is that we found that a certain gene is regulated in the vocal learning areas by vocalizing learned sounds, but not by other body movements or like walking, flying and so forth. And we find that the region of the genome that we think is controlling this specialized regulation of the gene in the vocal learning brain regions is a duplicated sequence that was not possible to assemble with the short read approaches. We only assembled it with the long read approaches. And now we have the specialized duplicated region and we can use it to genetically manipulate an animal who doesn't have the specialized gene regulation to now have specialized gene regulation like the vocal learners do. What is a reference genome to you? For me, a reference genome has been evolving. It began as I just need one individual genome to compare everything to to have as my base for the species. And you see people out there publishing. The genome of species X or Y or Z has now been done and complete and you go to it and you find it's incomplete. You go to it and you find that your favorite gene is rearranged because it's not really rearranged, it's just misassembled. And so then the reference genome became to me a genome that's highly accurate, that's ever free. And then I get to a more accurate assembly but the region I'm most interested in is missing because it's some repetitive region can't assemble. So now I think forever in the future I'm going to define a reference genome in the community. I think a large role eventually is that it has to be complete, no gaps. It has to be accurate so no errors in the sequences of the individual nucleotide, no errors in the structural variations. And it has to be representative of the species to a certain degree. Let's say representing 90% of the genome diversity in the species, which means you got to sequence many individuals. In your view, what are some of the barriers to achieving the goal of the perfect reference genome? In order to achieve this goal of this perfect reference for each species, with those three factors I mentioned, gapless, accurate and representative of the species, I think the scientific community or the funders and other groups need to bring people together of different backgrounds. We need to bring mathematicians together with the experimental biologists. We need to bring technology developers together with those two groups of people. And there's been an example of if you don't bring them together appropriately enough there's an example of what happens. What happened is we discovered in this vertebra genomes project that lots of errors in the existing genome assemblies is because genome assembly experts, the computer scientists, were trying to merge the two haplotypes together of a diploid genome into one, thinking there's not much of a big difference between mom and dad chromosomes. And by merging these two together, we're finding that they are creating errors in the genome assemblies that don't even exist in nature. And this is not because it wasn't doable. It was because you had computer scientists who didn't realize the biology of haplotypes well enough. And I think if they had partnered with biologists more closely in the beginning, we wouldn't have these kind of haplotype errors we have today, even with short-read assemblies. For you, why is it important to have reference genomes for all known biological species? That's right. Yeah, we need perfect assemblies for all of life to understand life, but also to preserve life because we're now in this current six mass extinction where the rate of extinction is 1,000 times higher according to some than it was before, like it was in the last mass extinction. We can't wait to get complete genomes of these species, so we need people to come together to do that now. And we don't know, at least for human society, we have no idea the knowledge that some of these species will bring to us until maybe sometime in the future when it's too late. What questions do you believe would all those reference genome assemblies allow you to answer? Good question, and I wrote an article on this once after we've done a bunch of bird genomes. What questions could you answer if you had all genomes of a particular class of animals or even of all of life? Because there are a lot of questions people don't even think about because they don't think this is possible. One question I think we would be able to address is to quantitatively determine what is a species? People don't realize a species concept is not really set in stone. And once we have genomes of all these different species, individuals of a population, so to speak, then we can determine, OK, should we really call this a separate species or not? The other thing we will be able to determine is infer the ancestral genome of a common group of species. And then take that genome, put it inside of an embryo, and try to recreate what that ancestral species was like. And then the other kind of question I think we will be able to address is, from my own favorite home project, is discover what other vocal learning species that are out there that potentially can imitate human speech. Because if we find we have found convergent changes in one group of species, like songbirds and humans, can we look for those changes in another species that no one's ever looked at yet for whether or not they can imitate and predict, OK, this species can talk to us as well. Are there any other areas of your work you think we should know about? Yes, so paper that we haven't discussed is a study that I worked on with members of my lab on trying to determine if mice have any resemblance of vocal learning abilities like humans do or songbirds do. And the answer that we and others in the community have discovered is no, they don't. They're not like a songbird, they're not like a human in their ability to imitate sounds and communicate in the way we do. But they're not as simple as people think. They have a rudimentary brain circuit and we think rudimentary genetic changes that we're seeing in songbirds, parrots, and humans. And that there is more of a continuum of this complex behavioral trait that gives us language that we're seeing in species, it's not black or white. And that gives me, and I think it should give other students out there greater appreciation of the diversity of phenotypes out there in the world, including things that the brain controls. But it also should mean that we should find a continuum of genetic changes in the genomes of all these species and more of a reason as to why we should sequence the genomes of many species. What advice would you give to a biology or genetic student, trainee, who is interested in tackling a risky question or subject? Now advice for a student is there should be nothing as, I don't think there is anything as a side project. What I tell them is go for the project that you really dream of, that you're fascinated by, because it's going to be just as much effort and work as a side project is. And don't be afraid to challenge the status quo, whether it be in genomics, or in physics, chemistry, or any area of science. It's okay, even if it seems like everybody is against you. If the evidence makes sense to you, then go with the evidence. That's the most important thing that I can tell you.