Ep 141: Vulnerability in science and in genomes (with Katie Letterhos)

Audio Block

Double-click here to upload or link to a .mp3. Learn more

What impact has the Trump administration had on biology and the scientific community? How do scientists study genomic adaptation and vulnerability?

On this episode we talk with Katie Lotterhos, Associate Professor in the Department of Marine and Environmental Sciences at Northeastern University. Katie is also the Secretary for the American Society of Naturalists, and she helped to coordinate and implement a survey of scientists to understand the impacts of the Trump administration’s policies on ecology, evolution, marine science and environmental science. In the first half of the episode, we talk with Katie about carrying out this survey, discuss some of the main themes of the results, and how the results can be used in the future. Then, we talk about Katie’s research where she uses oysters as a study system to understand the genomic basis of local adaptation and genomic vulnerability of populations to environmental change.

Cover art by Brianna Longo

• Episode Transcript

  Marty Martin  0:00  

  It's Fall 2025 and many of our listeners have, no doubt, been directly or indirectly impacted by the unprecedented attacks on science in the US, whether it's a loss of funding, the oversight of vaccines or support for universities, scientific research is being marginalized. While the reasons and justifications behind these efforts have been framed as attempts to reduce unnecessary spending, the motives are quite likely ideological. 

  Cameron Ghalambor  0:30  

  For example, not only is funding for scientific research being cut, federal scientists are being told they cannot use terms like climate change, sex or other words, with clear and agreed upon scientific definitions. Scientists in high level positions at the National Institutes of Health, the CDC and other top agencies have also been fired simply for doing their job. There's a lot of reason for concern that objective scientific research is being replaced with partisan and ideologically driven ideas by people from the political right.

  Marty Martin  1:05  

  But to be fair, scientific research has been under siege for decades. Earlier this year, Lawrence Krauss, professor emeritus at Arizona State University and president of the origins Project Foundation, published an edited volume called war on science. In that book, Lawrence and 38 colleagues wrote about the demise of scientific integrity in the US in the face of social activism.

  Cameron Ghalambor  1:26  

  The book was entitled War on Science because Lawrence and presumably the other authors felt that various groups across the country were compromising what science needs to remain a practically useful intellectual pursuit, even when the outcomes may not agree with a person's political or ideological views. 

  Marty Martin  1:45  

  Many chapters in that book raise great points. For instance, some highlight how a lack of public understanding of science as a process versus science as a fixed body of knowledge underpins many societal issues today.

  Cameron Ghalambor  1:57  

  Other chapters, though, are quite a bit more provocative and are written by people that many would recognize as fringe thinkers or extremely controversial. This is not to say that people I disagree with can't have good ideas. Another scientist's beliefs may not be aligned with my own, but I can still appreciate their work.

  Marty Martin  2:14  

   And even these more fraught chapters of Krauss's book, it's clear that the authors want some scientists to take partial blame for the current state of affairs. And one example they offer to make this point is the biological definition of sex. Any individual can identify as any sex or gender they prefer. But sex as a biological concept refers to differences in the size of gametes. If you want to hear more about this, we discussed the biology of sex with Hanna Kokko way back in Episode 17.

  Cameron Ghalambor  2:14  

  Long time listeners of big biology will know that we try not to dodge tough topics. From that episode on sex with Hanna Kokko to the genetics of human behavior with Kathryn Paige Harden, to the origins of the SARS-COV-2 virus with Alina Chan, we've taken a risk to cover things that we think are important and interesting. 

  Marty Martin  2:20  

  In fact, as a matter of policy, the only time we won't touch a topic is when we think the topic is too subjective, meaning not actually scientific, or the topic is just too far out of our wheelhouse to discuss meaningfully.

  Cameron Ghalambor  3:11  

  In the end, we enjoy doing this podcast because we think there's value in talking to other biologists about their research, and feel that one's political or ideological views should not bias the ultimate truths we seek to uncover.

  Marty Martin  3:24  

  It's exactly this feeling that inspired my colleagues and I to partner with the Heterodox Academy last December for an open conversation about Covid-19. We wanted to discuss a very complex, very personal and very controversial question: Did we handle the covid pandemic correctly? The conversations among more than a hundred medical doctors, virologists, epidemiologists, university administrators, local government representatives and lots of members of the public were very tough, but very worth it. And if you want to hear more about that event, we released a highlight reel as a big biology episode called Covid Conversations, just a few weeks back.

  Cameron Ghalambor  4:00  

  On today's show. Our goal is to confront directly just what recent, politically driven policy changes in the United States mean for science. Our guest today is Katie Lotterhos, a professor of conservation and genetics at Northeastern University. A few months back, Katie led a survey involving 14 different societies in ecology, evolution, conservation and organismal biology, about the damage that the Trump administration has done to the biological community in just nine months.

  Marty Martin  4:30  

  Most of us already know that the federal government is canceling long term partnerships like USAID, that for little costs, protected huge numbers of people from HIV and other infections.

  Cameron Ghalambor  4:40  

  They're also defunding research on messenger RNA vaccine technology at the same time when Katalin Karikó and Drew Weissman just won the Nobel Prize for their work in this area.

  Marty Martin  4:50  

  All the while, many of our leaders are trying to get the public to believe that safe vaccines are dangerous, and Tylenol during pregnancy causes autism, without providing any scientific support of the claims.

  Cameron Ghalambor  5:02  

  As you'll hear from Katie, these problems are just the tip of the iceberg. From stalled research, disrupted education and training programs to discontinued long term data collection, these policies are risking deep and permanent damage to most subdisciplines of biology.

  Marty Martin  5:18  

  The survey Katie led is a complex and detailed document, so we'll only scratch the surface today. We'll post a link to the results of the full survey in Substack so you can read it for yourself.

  Cameron Ghalambor  5:28  

  But in the spirit of being positive and productive, we wanted to devote equal time to Katie's amazing work on population genomics, the genomics of maladaptation and population vulnerability in oysters and other species. One of her main recent topics of interest is genomic vulnerability, how one uses genetic data to discern whether a population is at risk of extinction or other problems. 

  Marty Martin  5:51  

  Phenomena like adaptive divergence in genomic vulnerability are really tricky to study in nature, but Katie and colleagues have come up with some great guides on how to investigate these complex problems, such as what are called FST outliers. And these tools help them understand adaptation and evolution in the wild using just genetic data.

  Cameron Ghalambor  6:09  

  This chat with Katie reminded us that even though there are definitely problems in biology, we need to overcome and challenges we need to fix, modern biology is also more exciting and promising now than it has ever been before.

  Marty Martin  6:21  

  Daily, biologists are making important discoveries that have huge implications for our understanding of how the natural world operates and for how we improve our health and well being. 

  Cameron Ghalambor  6:30  

  We have to keep doing good work while fighting the good fight. There's just too much to discover and too much to lose if we let ignorance win.

  Marty Martin  6:37  

  And finally, like you, the less I hear about political topics on science podcasts, the better. We therefore promise that the rest of season eight, you can come to us for just biology goodness.

  Cameron Ghalambor  6:47  

  We have all sorts of great guests coming up this season, from Ruth Shaw on quantitative genetics to Jaap De Roode on animals medicating themselves, to Badru Mugerwa on big cat conservation.

  Marty Martin  6:59  

  And last thing before we start the show, remember that we're in fundraising mode.

  Cameron Ghalambor  7:02  

   Our goal is to get to 500 paid substack subscribers.

  Marty Martin  7:06  

  And remember, for every 20 new paid subscribers, we're going to randomly select a subscriber to receive a Big Biology t-shirt or a poster with a cover art from any of our episodes.

  Cameron Ghalambor  7:15  

  So for a chance to win, become a paid subscriber by going to big biology dot substack dot com and sign up to win. 

  Marty Martin  7:22  

  And once we get to 500 subscribers, that person will get a Big Biology sweatshirt and a large art print. 

  Cameron Ghalambor  7:28  

  And, of course, those prizes are in addition to the benefits that paid subscribers already get, such as access to whole episodes, episode debriefs and extra audio from our guests about their lives and hobbies. 

  Marty Martin  7:40  

  I'm Marty Martin

  Cameron Ghalambor  7:41  

  And I'm Cameron Ghalambor

  Marty Martin  7:41  

  And this is big biology.

  Marty Martin  7:52  

  Katie Lotterhos, Welcome to Big biology. 

  Katie Lotterhos  7:55  

  Thank you. Glad to be here. 

  Marty Martin  7:56  

  Yeah, we're really excited to have you on the show today for two reasons. In a bit, we want to get to the research going on in your lab. You've done a lot of exciting stuff on the genomics of maladaptation, genetic vulnerability of populations and many other topics involving oysters and other marine life. But first, we want to discuss the survey that you and others coordinated over the last months. To quote an email from our colleague, Dan Bolnick: "that survey was intended to quantify the effects of federal and state policies since January 20, 2025 on ecology, evolution, marine science and environmental science." So fourteen different American societies and organizations helped coordinate the survey. Can you maybe name some of those societies and then explain why they and you felt it was necessary to do the survey? Yeah.

  Katie Lotterhos  8:38  

  So the idea came and started with the American Society of Naturalists, and we are one of the oldest scientific societies in the US focused on biology. And we partner a lot with the Society for the Study of Evolution and the Society for systemic biologists, excuse me, Systematic Biologists. And we were concerned about a lot of the federal policies that have and other actions that were taken by the US Federal Government, starting on January 20, these actions resulted in a pretty large restructuring of the federal workforce. There were a lot of changes to historical norms. As you may have heard, there were a lot of executive orders. And we wanted to understand the impacts of all of these policies on our participants. So as we started to put together the survey, we started to reach out to other societies, including the Black Women in Ecology, Evolution and Marine Science, the Wildlife Society. Other societies started hearing about the survey. Society for Conservation Biology in North America, American genetics Association, Entomological Society of America, Ornithological Society, Women of Color in Ecological and Evolutionary Biology.  All partnered with us to complete the survey. So in total, we have 14 societies that the survey was sent out to. 

  Cameron Ghalambor  10:08  

  Wow. So Katie, you're currently the secretary for the American Society of Naturalists. So can you describe a little bit like, what your role was in coordinating and implementing this survey?

  Katie Lotterhos  10:25  

  So I took a leadership role in putting the survey together. We put together a list of questions that we wanted to ask. We shared that list with the societies. I put together a plan for the survey with the Institutional Review Board where we submitted that plan to Northeastern and that plan got approved. The importance of this step is going through a review to ensure that the privacy of the participants would be protected and that we would be in compliance with all federal laws when we distributed the survey. And so that plan includes not only what the survey questions are going to be, but how the data is going to be handled and how the privacy of participants is going to be protected. All of the little steps that need to happen to make a project happen are the same kinds of things you need to think about in a survey, and so I had the experience to lead this and bring this to life.

  Cameron Ghalambor  11:24  

  So you were you once the respondents came in the you, you were the one who actually, like compiled all of the results? 

  Katie Lotterhos  11:33  

  Correct. Yes, we did the survey through a system in our university that collects and compiles the results for us. So most of the results were compiled automatically in the system the quantitative type questions. But we asked an open-ended question, and we had over 700 responses to that question, and so where the real work for the survey came in was going through each of those individual responses and trying to summarize some of the themes and experiences that people were reporting. 

  Marty Martin  12:10  

  Well, we want to get to the themes, obviously. But first who responded? How many responses did you get?

  Katie Lotterhos  12:16  

  So we had around 1400 responses, around 58% of them were academic. So these would include faculty, postdocs and students. Around 42% were non-academic. So these are people working in industry, nonprofit, these included federal workers and as well as unemployed and retired workers as well. 

  Marty Martin  12:39  

  Yeah, okay. Can you maybe talk just briefly about, I mean, how you de-identified these data, the process of, sort of maintaining anonymity. I can't imagine that that was easy.

  Katie Lotterhos  12:51  

  No, that was one of the hardest parts. It's extremely important that, to us, that when we ran this survey that no one's identity would be compromised through reporting what their experiences were. And so when we think about ways, the kinds of information that people might be able to use to identify them, it's things like, you know what government branch they might work in, or what state they might work in, or what their pronouns are, or what their family does. You know, some people shared some very personal things. They shared the name of the species they worked on, the name of the project that they've been working on. A lot of those identifying details needed to be removed. In some cases, you know, all of the identifying details were removed because we were worried that any of them could lead to identifying that person. In other cases, the information was broad enough that we might be able to share, for example, like what federal agency they were reporting from. But in most cases, all of that information was removed. 

  Katie Lotterhos  14:04  

  I think it's also important to note that by removing this information, we lose a lot of the personality in the responses that people share. And I don't know how else to say this, but it like steals some of the thunder from what they're saying. Because, you know, those, those details, are part of the story, but in the way that people are vulnerable, right? Scientists are and students and consultants and everyone is vulnerable right now in the United States to having the funding that's supporting their science being taken away. You know, federal workers are vulnerable to having their livelihood be taken away. Immigrants are vulnerable to having their visas be taken away. We just couldn't risk anyone sharing their stories with us, you know, leading to an outcome like that. So we wanted to just be careful in how we were making sure that everything was appropriately deidentified. And so as the investigator, I went through and deidentified all the responses. Once I did that, I sent it to three peer reviewers, and so any responses that we had questions about were reviewed by all three of them, but at least every response was reviewed once by another person for this de-identification.

  Cameron Ghalambor  15:27  

  Wow. Okay, well, okay, so let's actually talk about some of these results. So on the website, looking at the results of the survey, you list kind of eight main takeaway messages, I guess, and we'll provide a link on our Substack site so listeners can go and see the actual results of the survey themselves. So I think, like the first major thing that seems to have happened due to these federal policy changes has been the disruption of research. So some of this research is focused on food security. Some of it's on like, flood mitigation, infectious disease, wildlife, all of these were sort of impacted by policy changes. Can you give a little more detail on, like, exactly, maybe some of the personal stories that came out of the survey on on this type of disruption, because at first glance, I think most reasonable people look at, you know, things like flood mitigation and food security as being kind of nonpartisan issues that, like everybody is on board for and would be sort of immune to these types of cuts, but that doesn't seem to have been the case.

  Katie Lotterhos  16:45  

  No, that doesn't seem to have been the case. Maybe the easiest way to answer this question is to read some of the responses and quote some of the responses that we have in the survey. Here's one from someone who says" "I work in wildlife research that protects both agriculture and public health. My team was awarded," and I'm going to say brackets in those de-identified areas: "[a specific type of bracket, funding to develop and test tools] to respond to a [specific disease] that threatens to devastate the pork industry. This work is vital to ensuring food security for the country. Despite having the expertise, infrastructure and urgency to act, we were stopped in our tracks. The funding freeze and administrative restrictions made it impossible to move forward. We had a clear mandate, but no ability to fulfill it. In many ways, this harm is irreparable." 

  Katie Lotterhos  17:45  

  Another example was from someone who was working with HIV. They said they're "at risk of causing a rapid rise in new infections of HIV if funding for prevention and treatment get cut." Another researcher, and I mean, it's across the board. Here's someone who was working on the impacts of sea level rise and pollution on salt marshes, and said that: "the local community where I work is engaged in this research because they saw how houses were protected by salt marshes during a devastating hurricane. The NOAA grants that funded much of my research have been cut. Moreover, this research examines the impact of two global change threats on salt salt marshes. We are not able to talk about or address global change issues in our grant proposals." So the you know, the types of science were really, really widespread in what people were reporting was being affected. 

  Marty Martin  18:44  

  Well, one other thing that I'm picking up on, but maybe you know, to broaden out, given the number of respondents you had, it doesn't seem like this was focused on any particular part of the country either. I mean, it doesn't sound like you had over represented samples coming in from this state or that state.

  Katie Lotterhos  18:59  

  We had responses from every state. They have those statistics somewhere, participants from all the scientific societies, every US state and most US territories. So there was, you know, some distribution according to the state's population size that we would expect. 

  Marty Martin  19:17  

  Sure yeah, New York and California, 

  Katie Lotterhos  19:20  

  Texas and Florida also were highly represented, but it was I was glad that we had responses from every state

  Marty Martin  19:30  

  And to just sort of continue on with these eight highlights. You gave a list of many different kinds of research affected. One of the other major things that the survey found is that there were pretty profound restrictions on free speech and travel, censorship of scientific terms, which you've already alluded to, concern about the ability of the government to comply with legal mandates, reports of bias to remove information from federal sources, and concerns about the future of long standing federal programs, with one example being the bird banding lab. So examples to share there?

  Katie Lotterhos  20:02  

  So many examples to share. Where to start with that. I think the you know, the restrictions on free speech and travel are really important to talk about, because as scientists, we need to have the freedom to use the terms to describe what we do, so that we can, we can have a common language and communication system, right? And I would like to read from one respondent who identified themselves as a federal worker who had received a banned word list. This wasn't the only respondent who said that they had received a banned word list, but this was one who gave us a full list of the words that were banned. They reported: "the banned words included scientific terms like and this is their quotes for the banned words, climate climate change, climate science, climate variability, global warming, carbon cycle, greenhouse gas emission, methane emission, clean energy, alternative energy, hydropower, geothermal, solar energy, solar power, photovoltaic, agrivoltaic, wind energy, nuclear energy, bioenergy, ethanol, electric vehicle, hydrogen vehicle, fuel cell, low emission vehicle, affordable home, low income housing, subsidized housing, housing affordability, runoff, microplastics, water pollution, air pollution, air pollution, ground pollution, pollution abatement, sediment remediation, PFAS, PFOA, PCB, non point source pollution, water treatment, water storage, water management, rural water, agricultural water, water conservation, water quality, clean water, safe drinking water." 

  Marty Martin  21:36  

  My gosh, what could you possibly write? It's not like most of those words are so broad. I mean, they're almost fairly inclusive.

  Katie Lotterhos  21:44  

  Other reported lists  of banned words including sex, gender and diversity. You know, these are scientific terms that we need to be able to use to communicate on a common ground with other scientists. We also saw restrictions on travel. Federal workers said their travel was restricted through not being able to use their purchase cards. But then others said their travel was banned even when it didn't cost the government. So they weren't allowed to go to meetings or even go to meet with collaborators, even if their collaborators were willing to pay for it. Another one said that they had travel restrictions so they couldn't perform their congressionally mandated monitoring for the National Park Service Inventory and Monitoring Program, and that was a program that's slated for dissolution under the restructuring plan right now. 

  Katie Lotterhos  22:42  

  Then there's, you know, there's another aspect to this which is this just the chilling effects that hearing about this has. And so there were some reports of people just self-censoring themselves in terms of well  trying not to use these terms because it could result in an automatic rejection of a grant proposal. 

  Katie Lotterhos  23:05  

  And then we have reports of immigrants being scared to leave the country because they're not sure if they're going to get back in on their visa, as well as scientists from other countries not wanting to come into the US because they're afraid of how they might be, what might happen if they're when they're here. Given some report of people coming in just to visit and being subject to what they feel is unfair treatment. So there was a few respondents who taught, who collaborate across the border with Canada, because, as we know, species don't know borders, and we need these kinds of collaborations to really get a full picture of how ecosystems are working and how species are operating in their ecosystems. So you know, there are reports of Canadian researchers not being comfortable to work in the US and moving experiments to Canada so they can continue their collaborations.

  Cameron Ghalambor  24:10  

  So, you know, I moved to Norway in 2020 but I still follow the news. It's kind of hard to get away from that. And so, you know, I was reading in the papers at the beginning of 2025 about Elon Musk and Doge and for public consumption the message was that, you know, there's government fraud and waste, and there needed to be these policy changes to eliminate all this wasteful money. But I think the other message that comes from the survey is that actually, in many cases, these cuts have maybe had the opposite effect, that there's actually been a decline in government efficiency. There's been a loss of institutional knowledge, which has like long term impacts. Decisions on funding have been delayed and and maybe that all of these, these cuts are actually, you know, causing more waste and less efficiency, rather than the opposite effect is, can you? Can you give some examples of maybe how the attempt to save money has actually ended up costing more money? 

  Katie Lotterhos  25:30  

  Yeah. I think one of the surprises for me from the survey is just how many types of collaborations there are. It's not just scientists at this academic institution and scientists at this other academic institution. It's federal scientists and people at the state level. Is people in nonprofits working with tribes, and universities and states and federal science. It's like there's just so many different ways that the industry and academia and states and federal governments and nonprofits are all working together, but it really depends on this, like federal cornerstone to operate, to keep those things going. And so here's one respondent who said: "majority of my research and outreach are funded by federal academic partnerships that leverage the resources of an academic institution to at least double." So I'm just going to pause here to explain that they're just trying to say, like the investment from the federal dollars is paying off double. So back to the quote. "So the cost effectiveness and practical impact of federal dollars invested in coastal aquaculture engineering, habitat restoration and flood resilience." They go on to say: "these partnerships are highly collaborative and nonpartisan, integrating research and resources across academia, federal/state/local agencies, nonprofit organizations, industry and the public and their effectiveness derives from the political neutrality with broad reach, which is achieved in part by being tied to a federal system of partnerships, abrupt loss of access to highly experienced federal personnel, at least one from the National Park Service, another from NOAA and another from the Environmental Protection Agency, whose advisory expertise was being actively used for projects involving biodiversity monitoring and close coastal flood protection. I have not observed a replacement for any of the three that makes up for their decades of knowledge. So these projects are now being advised with markedly less experience, which makes repetition of any past errors parentheses, and therefore waste of taxpayer funds and parentheses far more likely."

  Cameron Ghalambor  27:47  

  So I would imagine some people would say like, well, this sort of vacuum that's been created, of like federal funding could be filled by industry is there? I don't know if this was in the survey, but like, it would be great if industries, at least in some cases, could step up and fill the gap. But was there any good news like that? Of, like in the survey, of, 

  Katie Lotterhos  28:19  

  Let me quote the people from the industries who said that this is a bad idea. Here's a self identified contractor who highlighted how cuts to infrastructure restoration would prevent them from completing projects that would reduce damage from heavy flooding in their Midwest state. They said: "Irreplaceable harm has been caused because the public will not be able to depend on the accuracy of information in federal reports and publications. Already, we hear about data being created to support the desired research outcomes, which is illegal and violates every tenet of scientific inquiry." So the whole point of federal federally funded research is that it should be free of bias. And if we market this out to industry and to people that have concerted interests, or if the federal government starts to act for specific members of industry with concerted interests, we lose the very foundation of objective, rigorous science that we need to make good decisions for society.

  Marty Martin  29:27  

  Yeah, yeah. And, I mean, you know, industry, in some capacity, is and will have to step up for a little bit of this. And there's lucrative, you know, there's something positive for them in doing it. But the other dimension that I think isn't captured. I don't want to dwell on this. We've talked about it before. The reason the NSF was first conceived is to get the kind of research going that industry won't take the risks on that it's just absolutely too risky, or the payoff is too far in the future, or all of this kind of stuff, and that's just not something that's going to be supported. So another dimension of tricky.

  Katie Lotterhos  30:01  

  And I've been looking at different economic reports, and some of them are cited on our website that the survey is posted on, that suggests that for every dollar the federal government invests in basic, foundational, Discovery driven science, it has a return on investment of somewhere between two and three dollars like, who wouldn't want to make that kind of investment? Right? You know, the return is there, the evidence is there. But there seems to be this, this sentiment right now, that it's not worth it among policy makers.

  Marty Martin  30:37  

  I mean, this has sort of come up, but it was one of the other eight points of emphasis in the survey. It's not just about the current impacts on the people doing the research, the projects they're working on, the reliability of data. It's the sort of irreparable, long-term harm that this is causing. So I mean, how are people speaking to that? And I think one thing that we haven't touched on yet, I believe, is the next generation of scientists and the sort of consequences for training and education, for undergraduates,  graduate students.

  Katie Lotterhos  31:10  

  Those consequences and those concerns about an early-career bottleneck and career progress for early-career scientists right now was one of the main themes that we saw in the open-ended responses, you know, they were all concerned about, you know, direct cuts to training programs. So a lot of the grants that were canceled when Doge came through and started canceling grants were programs designed to train the next generation of scientists. This also led to many universities either directly cutting or scaling back on their graduate admissions because they were worried about the long term impacts of some of these policies. And so there was this domino effect outside of the federal sector that was predominantly affecting early career individuals and so, you know, we had responses from graduate students who were just concerned about, were they going to be able to find jobs, or what was going to happen after they graduated. Here's a quote from a someone who was probably a professor at a department said: "my entire Department hired just one graduate student this year, because although our funding has not yet been cut, the threat that it might be cut means that we cannot commit to paying graduate students in the future." So the threat of funding cuts has essentially the same effect as actual funding cuts, because it makes it impossible to plan or commit future expenditures.

  Marty Martin  32:52  

  Yeah, a lot of my colleagues told me their programs did not accept graduate students for that same reason. So this was a, I don't know if it's common, but it was well, more than ever should make sense.

  Katie Lotterhos  33:03  

  Yeah, and I think, I think this harm is irreparable when you look at it from the perspective of like we are taking these opportunities away from these young people at a time in their career when we could be inspiring them to figure out what they want to do with the rest of their life. And you know, at the same time, you know, we're cutting off that lifeline of that pipeline of knowledge that you know, these are the ways that we're translating our knowledge into the next generation of scientists. And when we take those opportunities away, we're also, you know, taking away that pipeline, yeah.

  Marty Martin  33:46  

  Yeah, so I mean when you first started having the conversations, was the expectation that enough members of the public would sort of rally to this? The problem now, Cam and I talked about this online is that every single day there's some five alarm fire

  Katie Lotterhos  34:01  

  Right

  Marty Martin  34:02  

  And so the media, you know, has its commercials to sell and the various different business things that it's obligated to do, but there's also so much out there to get to, and so many forms of it. So how do you, how do you get anybody's attention? And was the expectation that you get just a few people's attention, and then there's these educated mouthpieces that go to the legislators and make change?

  Katie Lotterhos  34:22  

  You know, honestly, the goal of the survey was more to give people a platform to share their stories. For people who wanted to do that, I wanted to provide them with a voice when they didn't feel comfortable speaking out themselves. If some good comes out of this, that would be great. Personally, I see it more as a snapshot in time that in 5 years or 10 years, we can look back to and say we have this evidence that things changed dramatically at this time point. And here's what it was like before. Here's what it was like after. You know, how do we get back to a place where people have the freedom to talk to their collaborators and have the freedom to use words to describe their science, and have the freedom to undergo peer review and have their plans for doing science, chosen on merit, rather than some political ideology. You know, I think it's more of a snapshot that I think we can look back to and use to justify that we need. In the future, we're going to need more resources to get back to where we were before.

  Cameron Ghalambor  35:47  

  So what's next?

  Katie Lotterhos  36:06  

  I think the best thing that listeners can do with this survey is to just share the link with people they know and share it with their representatives and governments at the state level and at the federal level, share it with the staff of those representatives. Get to know who those staff are. Talk to them about their concerns. Because honestly, I think a whisper campaign could be more powerful now than a big press hoopla.

  Cameron Ghalambor  36:35  

  So you mentioned, like, you know, politicians and their staff, like, did you sort of have them in mind when you completed the survey? Like, was that kind of like a target audience?

  Katie Lotterhos  36:48  

  Yeah, one of the interesting things I learned when I started attending communication workshops after January 20, I wanted to learn more about, you know, how to communicate science, how to talk to policy makers, was the importance of sharing individual surveys. And that was really a big motivator for the survey and to get this data out. I think, as a scientist, I often think like "Oh, my science needs to be, you know, peer reviewed and go through a very rigorous process before it can be used in policy". But actually, when you look at what's used in policy and what's used in court decisions, even, sometimes they're just citing blog posts. And so you know, doing a very carefully constructed technical report that you know underwent review with the societies who were part of it, but isn't published in like a big scientific journal, allowed us to get the data out very quickly without going through that very slow publication process, which we all know can take months and share it with the policy makers and the public and anyone who's interested to understand what these impacts are.

  Marty Martin  38:06  

  So one last question, Katie, before we switch over to, you know, a bit more happy topics. What was the sort of biggest surprise, or, you know, what did you take from this that I guess you didn't think would happen going in?

  Katie Lotterhos  38:20  

  I think the thing that surprised me, personally, the most was just how many ways the federal government facilitates science. And some of it isn't like really require a lot of monetary resources. It's just, you know, having the people with knowledge in the right positions to give advice and to collaborate and to, you know, have that knowledge base in the federal government and provide objective information that industry can use and and farmers can use, and you know, oyster growers can use to make decisions about how to conduct their livelihoods. So that was a surprise for me. I think the other, it's not a surprise, but the other thing that it made me reflect on is that, you know, we really need freedom for science to work. You know, having the freedom to exercise values like objectivity and rigor, the freedom to describe natural phenomena with the correct terminology, you know, the freedom to question and debate what we believe to be true. Like these are freedoms that we need for science to operate and to advance knowledge. And they're not freedoms for scientists alone, either, like, they're just fundamental requirements for a society that values truth and, you know, the common good. 

  Marty Martin  39:52  

  Yeah, absolutely. 

  Katie Lotterhos  39:55  

  The one quote that I would that I really think resonated the most with me out of all of the quotes, was someone who said, quote: "The environment is part of the American identity". And I think that is so true when you look at people across the political spectrums, we all have our different ideas of you know, what's fun for us, but everybody enjoys the outdoors and being outside. And if it's going to the beach or going in the woods, you know, it is something that I think a lot of Americans value and share and have in common, so maybe that eventually will be the common ground that brings us back together.

  Cameron Ghalambor  40:51  

  Should we transition to something more, more uplifting, like all the cool research you're doing? Yeah, I want to just start off by saying, I'm a huge fan of your research. I find your papers to be extremely thoughtful, and so I can see how you brought that same thoughtfulness to the survey. To start off with, I want to talk about a topic that I you've you've thought a lot about, which is the genomic basis of local adaptation, I guess, for context, you know, for our listeners, you know, historically, you know, this is one of the big questions in in evolutionary biology, evolutionary ecology, of how do populations become locally adapted, which traits are under selection. You know, is fitness higher in your home environment versus in an away environment. And now, with the advent of DNA sequencing technology, we can actually identify regions of the genome, specific genes or markers, that are more differentiated than we would expect given some neutral processes operating on the genome. And this is, you know, you can't open up a scientific journal in evolutionary biology without seeing the application of these methods in some way. And so you've really thought a lot about some of the assumptions that go into these, these type of methods, and I guess in particular, we should introduce the term of the F statistic and these FST outliers. So could you really briefly, like, for you know, if you were having to explain this to your family or something. Give us a kind of just an overview of what is an FST outlier and why is it such a powerful method for detecting selection on the genome?

  Katie Lotterhos  42:55  

  Well, my family will probably listen to this, so I can ask them how my explanation was when I'm done, but it's basically the idea that, well, let's back up, so local adaptation is basically the idea that within a species, you know, there's some populations that are going to be better adapted to one kind of environment, and others that are going to be better adapted to another kind of environment. So there's a lot of diversity within species. I work on oysters, so I think a lot about oyster populations that are adapted to like lower salinity environments or higher salinity environments or higher disease environments. And so those adaptations can have a genetic basis. And so to understand what that genetic basis is we can sequence their genomes, and then we can essentially scan across the genome, and we can look for sites in the genome where the genetics is different in those two populations. So we're basically letting the evolutionary history do the work for us that the evolution has already selected for these specific genetic markers that allow individuals to survive in that specific environment. And so simply by collecting individuals from those different populations and scanning their genomes, we can look for these genetic regions where places diverge. And so FST is this statistic where it essentially measures the amount of genetic differentiation at a specific place in the genome. So we can scan along and measure FST, and at places where the populations are genetically different, there's a higher FST, and that is potentially a place in the genome that's under selection by the different environments.

  Marty Martin  44:42  

  Okay, so I think, you know, likewise, every journal of evolutionary biology, and maybe a good fraction of any journals with biology, at this point, you're going to find what are called Manhattan plots right here, which is the visual depiction of these FST outliers. When I first saw them, you know, Cam was sort of surprised to hear me say that, "Wow, I thought that was cool". He thinks I'm quite skeptical about some of this sort of stuff, you know, take away my evolutionary biology playing card. But, you know, as an organismal biologist, I do think about things as integrated networks. And we've talked a lot on the show, and I know you've thought a lot about the roles of epistasis and pleiotropy, and you know, basically just that genes' effects on phenotypes are super complicated. So whereas these Manhattan plots and FST outliers work, I mean in a nutshell, and I'm asking you a very simple question on where, in a nutshell, how do we think about these complex epistatic, pleiotropic effects in relation to outliers?

  Katie Lotterhos  45:41  

  I haven't looked at epistasis specifically, but I have looked at pleiotropy in some of my research. And I should back up to say, the way that we evaluate some of these tests and how accurate you know they are, they're giving us results is by doing computer simulations. And so when we simulate things on the computer, we know the truth. We know which genetic markers are under selection, because we're telling the computer. We're giving those instructions to the computer. And so then we can run that output from the simulation through the method and see how accurate the method is. And so we've done some of these simulations with pleiotropy. And we did find that when we added pleiotropy, the power, or like, the ability of the methods to accurately detect the genetic markers involved in local adaptation, did decline. And so, you know, it's no question that these methods are missing, probably missing, like some of the more complex interactions that we know occur in genomes. But nevertheless, they have also been extremely successful at identifying a lot of the genetic markers that are involved in adaptation, and this has important applications in selective breeding across a lot of different fields.

  Cameron Ghalambor  47:10  

  Yeah, so, you know, I first became familiar with some of the papers that you wrote with Mike Whitlock now, probably over 10 years ago, and I think now they've become maybe part of the conversation of like things that you have to do when you when you start to analyze your data and look for these FST outliers. But one of the themes that you've kind of talked a lot about, is the the importance of demographic history and and population structure, and how that can, I guess, for lack of a better term, maybe give false positives and give you the impression that a region of the genome might be under selection when it actually isn't. Can you talk a little bit about how, how that plays out, like in maybe in an example from the oysters, or something that maybe listeners can relate to?

  Katie Lotterhos  48:05  

  Yeah, this is a major challenge in applying these tests. So just to back up a little bit, the we're talking about looking for genetic differences between populations that are due by adaptation, but there can also be genetic differences that evolve between populations that are coming from other processes operating. So there's, like, a lot of randomness in the genetics of populations. There's randomness and what's passed down from generation to generation. There can be, like, random hurricanes or volcano explosions, you know. And then we also have genetic exchange through the movement of individuals. So, you know, in trees, pollen can travel thousands of miles. In oysters, their larvae can travel long distances in the ocean. So these all contribute to what we call population structure. And so, you know, when we're doing the assert, these scans, where we look across a genome, we find a lot of sites that are different among populations, but most of them are probably caused by these like random events I was talking about. So if we don't account for that structure, we find a lot of genetic markers that are different, but probably like not that many of them are involved in adaptation. 

  Katie Lotterhos  49:21  

  And so the major challenge is, how do we like subtract away the signal of population structure so that we can more accurately identify the genetic markers? And that's where my research comes in. I've looked a lot at like these different methods and how they correct for a structure, and trying to identify the best ways to apply these methods in in different scenarios, and this is what I call the Catch-22 of population genomics. So like, if we don't correct for structure, we get a lot of false signals, genetic markers that look like they're involved in adaptation, but they're really not. But we've also shown that under some conditions, when you correct for structure. Or can actually dampen the signal of adaptation that you're looking for an example of that. So the oysters that I work on go from Texas to Nova Scotia, which is basically like the entire East Coast of the United States, and in the gulf between Texas and Florida, they have quite a lot of different genetic markers. I guess you could say that on the East Coast, because you got Florida there and it doesn't, it kind of prevents the oyster larvae from really traveling around. So there's differences, but they're probably caused mostly by this, you know, lack of gene flow or genetic exchange between the two locations. But there's also a lot of environmental differences. So the Gulf is a lot warmer than it is on the East Coast, and so if we want to use one of these tests to look for, let's say, like differences that might be due to temperature adaptation, being adapted to more warmer environments in the Gulf by correcting for this structure, which is quite strong between the two locations, we might actually be subtracting away that signal that we're looking for related to temperature. And so, you know, there's this, this Catch-22.

  Marty Martin  51:19  

  So we have, well, I am speaking out of ignorance, So correct me if I'm wrong. But in principle, we have lots and lots of FST estimates and lots and lots of things. We had Mike Lynch on the show just a little while ago, and we talked many about many things, but one of them was effective population size. So without getting into the weeds, listeners can go and hear all of the reasons that that's such is an amazing thing. But in light of effective population size, is there sort of something systematic that you might expect where it's harder to find genomic evidence of adaptation, because effective population size is inherently smaller, and you know this great big, long lived species, for example?

  Katie Lotterhos  52:02  

  Potentially, I think there's a there's, I think about this from a first principles, as well as a statistical perspective. So when effective population size is small, those random processes that I was talking about, like, really dominate the population, right? And so that makes it harder for species to adapt. And so in that case, the question is, if we don't discover something, is it because it's not there, or is it because the statistical tests that we're using doesn't work? And so I might not have a good answer for your question, other than sometimes, as a scientist like we need to think about things from this first principles perspective, but also from this statistical perspective of, is there some property of the data that's preventing us from getting the signal, or is the actual lack of signal a property of the data, if that makes sense?

  Marty Martin  53:01  

  Yeah, yeah. No, that makes sense. I guess you know, it just makes me wonder. But because there are so many these FST estimates out there, it seems like, and this is my ignorance, it seems like there might be an opportunity to reconcile how much of this is statistical problems versus interesting biology influencing the way that you know genomes can adapt 

  Katie Lotterhos  53:22  

  Absolutely. Yeah. And actually, it's funny, because one of the methods that I've worked on called OutFLANK, is one of these FST outlier methods. It's more conservative with the than the other methods, which means it doesn't return a lot of outliers. And every once in a while, I'll get someone emailing like, "Hey, your method didn't find any outliers in my data". And I'll be like, "well, maybe there aren't any outliers in your data".

  Marty Martin  53:46  

  You know, no guarantees. That's not how this works.

  Katie Lotterhos  53:49  

  Money back. So there could be some publication bias there as well, and that we tend to publish things that have signals.

  Cameron Ghalambor  53:57  

  Yeah. So, so Katie, I'm also interested in, you know, like you mentioned, salinity, and I've become interested also in salinity these past few years. And we know a lot about the not just the individual genes, but actually the pathways that those genes are part of. And so, you know, as part of the physiology of the organism to maintain ion homeostasis in response to different, you know, salinity challenges. We have a pretty good idea of, like, the network of genes that get upregulated in response to these kinds of salinity challenges. And I'm curious, like, when we think about, like, FST outliers as, like, you know, structural differences in the genome, but then also for traits that are, you know, plastic and being upregulated and down regulated in response to the environment. Obviously, these kinds of methods are going to be biased towards picking up genes that have very large effects. What are your thoughts on that? Like, do you have you? Do you think about these, these kinds of challenges?

  Katie Lotterhos  55:12  

  Absolutely. Yeah, yeah. I actually, I think it's interesting, because when you look in the literature, some people seem very interested in, like doing an FST outlier test and then doing like an RNA-Seq, where they're looking for gene regulation changes, right? And then they're looking at the overlap between those two things. But what evolutionary theory says that there should be overlap between those two things, right? We, on one hand, if some genetic markers genetically adapting, then it's changing the DNA sequence. If that's happening, it may not have to have a plastic response, or a response to the environment on top of that genetic exchange. When we look at theory, things that are having a more plastic response are not necessarily going to adapt, because that plastic response is already there, and so we might not expect to see genetic change at those places that are plastically responding. So the two approaches are complementary, but that doesn't necessarily mean that we expect to find the same signals when we're taking those two different approaches to studying adaptation. And the other piece of that too Cam is, you know, some of these mechanisms that that species are using to adapt to environments can be involve, like thousands of genes, each making really small contributions. And these are traits that we call polygenic and these methods are designed to detect these loci of large effect that are showing really big genetic differences between populations. But if it's a trait that has a lot of sites across the genome that are all making small contributions, we can still see strong patterns of adaptation. And some of my work has shown this, and others have shown this. Even though there's only really small changes in the frequency differences between populations, it's kind of like there's a lot of small changes that build up, and so we can't detect those using like these traditional FST outlier tests.

  Marty Martin  57:27  

  So we had Scott Edwards on last season, and that was specifically about structural variants in the genome. You know, big, big variation. Is this something that you're pursuing now?

  Katie Lotterhos  57:38  

  Yeah, I love inversions. I've been doing a little bit of work with inversions. So an inversion is a place in the genome where the DNA has essentially been flipped around. And when that happens, there's not a lot of recombination that happens between those two variants. And then, as a result, the amount of genetic exchange is also reduced in that area of the genome. And then, as a result of that, it allows that those regions to diverge. And so when we look across the tree of life, we find a lot of species that are locally adapted being mapped to these really large inversions. And few years ago, my graduate student, Sara Schaal, published a theory paper with me where we did simulations. We called it inversion invasions. So we basically did a very basic quantitative genetic model, which is just basically a trait, and there's a bunch of genetic markers that have additive effects on that trait, and it's under selection to different environments, and we let inversion mutations invade the simulations and just just saw what happened. And what we found in that study was that the inversions that were involved in adaptation tended to be larger and older than all of the other inversion mutations in the genome, which is in line with what we see in nature. But two other really interesting things came out of that study. One was, when we look at a lot of these systems in nature, the inversions that seem to be involved are happening under high gene flow. So for example, my lab has been looking at Atlantic cod. They have a migratory and more local populations in Iceland and where you are in Norway, Cam. And the genetic basis for this has been mapped to four large inversions. 

  Katie Lotterhos  59:39  

  When we look at historical inversion theory, it's really focused on these, like low gene flow scenarios for one inversion. And so the question we were able to address in our simulations was like, why do we see multiple inversions involved in adaptation and genomes? And, in our simulations, this happened. And under high gene flow, when the genetic architecture was highly polygenic, actually, so there was lots of loci of really small effects. And when an inversion mutated, it tended to capture a few of these loci. And then over time, the genetic architecture became concentrated in this inversion with more and more loci of small effects in our simulation. 

  Katie Lotterhos  1:00:26  

  And the other thing that was really interesting in our simulations was that overlapping inversions evolved, and so these aren't what I would call intra-chromosomal overlaps. It wasn't an inversion on top of inversion on the same chromosome, it would be like an inversion would evolve in one population and be involved in adaptation, and then an inversion in another area of the genome in another population would become established. It's in the same part of the genetic map, but it's in a different individual in a different population. 

  Marty Martin  1:01:01  

  Wow. 

  Katie Lotterhos  1:01:02  

  And we couldn't see these in the population genetic like FST signals in the simulations, which I think opens the possibility that maybe they're out there in real data. And I just wrote a grant proposal to look for them, so we'll see if that gets funded.

  Marty Martin  1:01:19  

  Oh cool. Okay, to be continued, just one quick follow up on that Katie. So how do you know, without going into the methodological weeds, I don't want to ask you to go super far down that road, but how do you know these are inversions, as opposed to just great big expanses of sequence?

  Katie Lotterhos  1:01:34  

  Yeah, so when we do the whole genome sequencing, what happens in an inversion is that a read is going along and then the rest of it gets mapped way over to another place in the genome, and it's going in the opposite direction. And so we can actually look at the way that the read that came from an individual is mapping onto our reference to tell that it's an inversion,

  Marty Martin  1:02:02  

  Yeah okay, so it's specific to the sequence that it used to be, as opposed just the length of something else popping in.

  Katie Lotterhos  1:02:09  

  You discover it from like comparing the sequence of the individual that you're sequencing to a reference, essentially yeah.

  Cameron Ghalambor  1:02:16  

  Yeah I think the fact that these inversions, and I guess, you know, structural variants more broadly, are segregating in populations, is been probably vastly underestimated historically. And it does seem like that's the you know, coming fashion, so to speak. You know, especially with now having long read sequencing. And, yeah, we talked with Scott Edwards about pan genomes. And rather than going to a reference, just comparing against other individuals to see, you know what's going on. 

  Cameron Ghalambor  1:02:16  

  Okay, I have another question for you. It's kind of related to a manuscript that we have kind of in revision, but, but also, but I'm going to keep it kind of vague, just to kind of generalize it. So let's say you have you talked about gene flow, and so that kind of type of genetic exchange is going to, sort of usually we think of as creating this sort of background of, you know, neutral background that we're trying to pick up the FST outliers against. But like in the case of, say, organisms that move and can choose the habitats in which they settle in. So let's say you have variation in the population of oysters for their ability to tolerate salinity, and these oysters can move to some degree. Maybe they have little feet or little they can swim, I don't know, but-

  Katie Lotterhos  1:02:16  

  They glue themselves to the rocks. 

  Cameron Ghalambor  1:02:16  

  I know oysters maybe are not the best. Let's switch to cod. 

  Katie Lotterhos  1:02:16  

  Okay

  Cameron Ghalambor  1:02:21  

  Let's switch to cod. I realized very quickly that I had chosen the wrong organism. But let's say they're fish and so you have, you have segregating variation in the population for salinity tolerance. And, you know, some individuals go to the salty area, some individuals go to the freshwater area. You do your FST outlier test, and sure enough, beautiful differences between habitats. How do you separate selection? Because I would interpret that at first pass as divergent selection has generated these differences from just habitat selection and the behavior of the organisms.

  Katie Lotterhos  1:04:49  

  I'm confused because this are we assuming habitat selection doesn't have a genetic basis or it does?

  Cameron Ghalambor  1:04:54  

  Yeah, I guess I would say it doesn't, yeah. So like, it's just be, you know, like you just match your phenotype-genotype to an environment.

  Katie Lotterhos  1:05:03  

  I'm still confused . So, okay, let me back up. So if habitat selection and has a genetic basis and salinity tolerance does not, and we more we we look at these genetic differences, then they might look like they're due to salinity tolerance, but they're due to habitat selection. Is that what you're asking?

  Cameron Ghalambor  1:05:25  

  Yeah, basically, like you see, you see genetic differences between populations and different habitats, but it's not because there was a single population that was under divergent selection.

  Katie Lotterhos  1:05:36  

  I see what you're getting at. No, I think this is an interesting challenge, and this is the challenge of identifying the selective force, right? Because we have a lot of correlated environments out there, and in my own research, this is definitely a challenge. In oysters, it's not just differences in salinity. The salinity gradient also is very highly correlated with the disease and micro parasites that we find in oysters. And so lower salinity sites are stressful for oysters because they're lower salinity, but there's a lot fewer of their parasites that survive at lower salinity. And so it's a harbor from parasitism. Where at the higher salinity sites, there's three or four different parasites that we find in higher frequencies in the oysters, we find a couple of protozoan parasites, as well as these pea crabs that like to settle in oysters. They're like these little, tiny crabs that go into the oysters as larvae, and they grow up to be about the size of your fingernail, and they have, like, make a little home inside the oyster shell. And so when you open up the oysters, the gills, which are what the oysters use to filter feed, are substantially reduced to the point where they have almost no gill, because that's where the pea crab likes to live. So we have this problem of correlated environments. And so I think when we do these tests, you know, it brings in the importance of the biology and the need for experimentation, because you can't always isolate the true selective force in these really complex, multivariate environments that organisms live in.

  Marty Martin  1:07:28  

  Okay, so let's apply these FST outliers in kind of a context that I think is of interest to you and lots of people. There's a paper that you wrote not too long ago. I think it was in Evolution Letters on the concept of genomic vulnerability. And I think a lot of that motivation has to do with mitigation of climate change and conservation in general. But there's a tool that you advocate there you sort of focus on, called genomic offset. So perhaps describe that, relate it to these FST outliers, and connect it to genomic vulnerability. A lot of complicated terms.

  Katie Lotterhos  1:08:02  

  Yeah, so these kinds of studies that we've been talking about are looking across landscapes, they're looking across species ranges, and they're looking for genetic differences in populations and how those relate to the environment. There's parallel studies to FST outliers called genetic-environment associations. They allow you to hone in a little bit more on the specific environmental variables that are related to genetic differences, but they produce the same kind of output as FST outlier tests. And so the question is, you know, can we use these contemporary genetic patterns across environmental gradients to predict the responses of populations within a species to climate change? And that's where this idea of a genomic offset comes in. It's a measurement of how maladapted a population will be to an environmental change. And so the idea is, if a population possesses the genetic sequences that will allow it to be adapted to the environment in the future, then the genetic offset would be zero. But if it requires a lot of genetic change to be adapted to the environment in the future, then the genetic offset would be a large number. And so in this way, it's a forecast of how the population is going to respond to a future environment.

  Cameron Ghalambor  1:09:29  

  Yeah, yeah. So Marty and I were in Barcelona a few months ago for the European Evolution Meetings, the ESEB meetings, and I saw a lot of talks using genomic offset. 

  Katie Lotterhos  1:09:44  

  It is a rapidly growing subs field.

  Cameron Ghalambor  1:09:46  

  Very much so, and, and so, you know, I know you've reminded us, and Marty mentioned this Evolution Letters paper. And I think there's a special issue of the American Naturalist that's coming out. And so you've reminded us a lot about the assumptions that go into using these kinds of methods, and so I guess one of the one of those assumptions has to do with the degree to which populations are locally adapted or maladapted to their different environments, and the degree to which you can infer adaptation versus maladaptation, simply by looking at this like subset of loci. And so how good of an assumption is that, like, you know, Marty alluded earlier, to the sort of the complexity you know of going from genes to phenotypes, and so is that if a graduate student is using these methods, should they be concerned about that assumption?

  Katie Lotterhos  1:10:51  

  I think they should absolutely be concerned about that assumption. I think really understanding how well populations are adapted to their local environments is not easy to do. To really do it well, you need to do these really large scale reciprocal transplant experiments, where you like, take a whole bunch of different populations and you transplant them to every single location, and you measure the performance of all of those individuals. When we're doing these kinds of landscape association type tests with genetics and environment and making forecasts from them, we're using existing patterns to make a prediction. And there's a lot of assumptions that go into that, and like you said, one of them is that there's essentially very strong patterns of local adaptation, and like the work we've done before, we've also used simulations to evaluate some of these genetic offset methods. And we found in those simulations that the performance of the methods can vary greatly depending on, again, the species demography, the evolutionary history, that genetic architecture. And we did discover that the accuracy of genetic offsets increases with the amount of local adaptation in the population, but that also the accuracy decreases as the environmental change becomes more novel. And so that means as the environment changes beyond what the species has historically experienced within its range. So these are two really important caveats right now, of using these offsets, and I think the number of studies that is clear where they're extrapolating environmental change to, and how novel that environmental change is in the future compared to what exists in the species. This isn't something that's clearly reported right now in a lot of these genomic offset studies.

  Marty Martin  1:12:54  

  Can you say more Katie about the novelty? I mean, it sounds like that's novelty in degree. You know, things are warm or more variable. It is not novel in kind. 

  Katie Lotterhos  1:13:04  

  It's a multivariate novelty. So it's like saying, you know, here's the climate envelope, here's the salinity and the temperature. And in the future, you know, the salinity and temperature is going to be outside of this range that we're training the model on, right? That's an extrapolation.

  Cameron Ghalambor  1:13:24  

  So I find that really interesting. So the envelope is the data that goes into determining the size of the envelope based on sort of the average kind of temperature? One of the, one of the things that I worry about a lot is that, you know, at least the systems I work on, environments fluctuate widely from year to year and and so what's adaptive, you know, a genotype that you would say is locally adapted at time x is not necessarily the same genotype that's adapted at time x plus one?

  Katie Lotterhos  1:14:03  

  Yeah this is a good question. So let's, let's talk a little bit about how the methods work. And then I think there's a philosophical question here, of like, how do we quantify the environment?

  Marty Martin  1:14:17  

  It's a trivial thing. It's very easy. Yeah, yeah.

  Katie Lotterhos  1:14:21  

  Yeah really, sounds easy. How the methods work, what makes them unique is how they're integrating this genomic and environmental data, you know, typically, into some kind of machine learning model, and that model's learning from the data how the genetic patterns are changing across environmental gradients. And so it's combining signals across multiple genetic markers and multiple environmental variables to make a prediction of how much genetic change occurs per given environmental change. And so, you know, they're being trained on these contemporary patterns. One simply just plugs in their data and plugs in what the environment is going to be in the future, and the model outputs a genomic offset. And so you know the environments that you're putting into it are going to determine, you know what the offset is calculating. So if you're missing something important, like disease or parasitism, you know, that's not going to be included in your forecast. 

  Katie Lotterhos  1:15:26  

  The second question is when the models are trained, you put in something like, like temperature, right? And so it could be the average temperature for each population. You could put something in, like the standard deviation in temperature to try to get at variability. But then, like, what does that mean? You know, the standard deviation for temperature for a really cold site could be the same as the standard deviation in temperature for a really warm site. You could say, like, "Okay, well, maybe I'll just put in the maximum temperature, or the average maximum temperature". Well, if you want to get into the statistical weeds of that, which my lab has had to do recently, because we have in our oyster study data sets from like, a bunch of different data sources, there's not like in the terrestrial realm, it's easy to get things like temperature data standardized, more or less. But in the marine realm, there's just so many different buoys. Some of them are state, some of them are federal, some of them are institutional, and we're putting all this data together, and they all have different frequencies right of measurement. And it turns out, if you take the maximum, if you have two sites that are exactly the same in their temperature, but one of them has more frequent than the other, and you take the maximum of both of those, average it every per some unit time. The one that has more frequent monitoring is going to have a higher maximum than the one that doesn't. And this just comes from statistical sampling, right? And so, you know, there's, there's a lot of questions I have about, like, how do we accurately quantify the environment to train these models? And that's an important question that investigators applying these models have to consider.

  Marty Martin  1:17:16  

  Well, I, you know, I think we all love the idea of genomic offsets to describe vulnerability. But can you describe, can you name a sort of case study where forecasts have been effective, or everyone expects them to be effective?

  Katie Lotterhos  1:17:34  

  Yeah, so there's, there's a few empirical evaluations right now of genomic offsets, I would say the majority of papers are producing a forecast but not testing it. And I think this question of evaluation is really important, you know, because we need to validate the models, we need to be able to tell if they're making predictions that are accurate. So the way that people have approached this, most of these papers have occurred in trees. They basically collected trees from different populations and raised them in common gardens. Measured the fitness of the trees from different populations in that common garden. When you move them from one location to another, you have an environmental change. So that's what's used to calculate the genomic offset. And then you can compare what you see in the field to what the genomic offset prediction was. These are the kinds of studies that we need right now to really evaluate these methods. And so there's been about somewhere between, you know, six to ten of these studies that are published, some of them are coming out of Steve Keller, Matt Fitzpatrick's labs. Brandon Lind was a former postdoc in my lab who published one, and we have a few more that are going to be published in the special issue of American Naturalist.

  Katie Lotterhos  1:18:45  

  And across all of those, we see a lot of context dependencies. It can be dependent on the genetic markers that go into the offset calculations, the populations that go into the offset calculation, the site, the common garden site, where the trees are grown. We're spearheading the first tests of genomic offsets in marine species, and so one of our tests is in oysters. We've been growing oysters at high and low salinity sites in the Chesapeake Bay, and another test is in collaboration with Marlene Jahnke at the Gothenburg University in Sweden, where we've been looking at sea grass across the Baltic Sea and predictions for sea grass and raising them in in current and future Baltic environments. So you know, so far, our preliminary data don't find that genomic offsets are predicting these specific species really well in oysters. We're finding that oysters from more distant locations tend to do better than oysters local to our field sites, or at least as well as them. And it seems that oysters from the south, from more southern locations that are warmer and have higher salinity, tend to do better at both of our field sites. We think that might be because of a recent history of warming at our site. So local populations are already kind of lagging behind climate change and distant populations are tending to do better. So there's a lot of complexities in taking these patterns, these contemporary patterns of genetics and environments, and trying to make predictions. And it's probably going to work in some species and probably not in others. And so we need to figure out, you know, what are the conditions under which these predictions are accurate?

  Marty Martin  1:20:43  

  Yeah, yeah. Something you said there, Katie, I have a follow up. Are people trying? Or do you have plans to use transcriptomic data at all? Or is everything sequence based?

  Katie Lotterhos  1:20:57  

  This is an excellent question. I mean, right now, all of the methods are sequence based. But when you look at the kind of AI medical literature, they're really integrating a lot of different data sets, transcriptomic, epigenomic, microbial biology, like, I think this is a philosophical question of like, what kind of data do we actually need to make an accurate prediction, right? 

  Marty Martin  1:21:22  

  Yeah, yeah. Well, so I think the genomic vulnerability, like, that's a really compelling idea to me, but my genomic tends to be transcriptomic. Because I think, well I am, more of a physiologist than an evolutionary geneticist.

  Katie Lotterhos  1:21:36  

  There's no reason you wouldn't be able to use the same data in the same way. I think, I think the challenge with transcriptomic data is that it's so time dependent 

  Marty Martin  1:21:36  

  Oh yes

  Katie Lotterhos  1:21:37  

  And, you know, stage dependent and tissue dependent, right? 

  Marty Martin  1:21:51  

  And tissue and age and sex yeah.

  Katie Lotterhos  1:21:52  

  Yeah. And so the appeal of the genomic offset is maybe that it's, well, the DNA is there, and it's in all the tissues. It's the same. But it is a really cool idea.

  Cameron Ghalambor  1:22:07  

  Yeah, no, I Well, and the question that I was going to ask is what you just touched on, which is the, I think the bigger problem of context dependency. And I guess, yeah, this is a philosophical question, but you know, it's easy to go out and get a bunch of DNA data now and sequence it and get sequence data and feed it into a model and get an answer. It's a lot harder to quantify the environment, as you were just describing, and quantify phenotypes, because, you know, given the, again, the complexity of, you know, the architecture underlying different traits, the interactions of those traits. I mean, I'm, I'm also guilty of, you know, using like, thermal tolerance data to make conclusions about vulnerability in the future, but at least those traits are a little bit closer to fitness than, you know, some allelic change. And so, I guess, yeah, just to kind of maybe wrap up, you know, if you're if somebody who's listening to this episode is like a graduate student and is really interested in the idea of genomic offset, like, if you had to give them a very quick, like, sort of bit of advice before they they jump in. Would you Is there some really bit, short bit of advice you could give, or is it, you know, be prepared for a lot of-

  Marty Martin  1:23:37  

  Good luck

  Cameron Ghalambor  1:23:39  

  -a lot of challenges?

  Katie Lotterhos  1:23:42  

  Do an experiment. Yeah, just design it in a way that you can test what the predictions are, because you will learn from that. And even if the method doesn't work, we have a dataset that we can add to the growing number of data sets that we can use to develop methods. And you know, once, once we have enough of these datasets in place, we could actually develop some kind of system to rank methods and test them. And these are approaches the data scientists use and people developing AI algorithms use to like, determine what the best algorithm is to solve specific problems. And so, you know, designing your study in a way that you can do an experiment to test a forecast is really, would be really cool. 

  Cameron Ghalambor  1:24:33  

  Yeah, no, that's, I think that's great advice.

  Marty Martin  1:24:35  

  Okay, we really appreciate your time. I gotta say, I think this latter part of our conversation was a lot more uplifting than the first part, but the first part was important to cover. We want to make sure to give you some space to, you know, mention anything about your research or the survey or future plans that we didn't, didn't prompt you on. Is there anything else that you wanted to say?

  Katie Lotterhos  1:24:54  

  I wanted to say one other philosophical thing about these forecasts, which is, if somebody finds a forecast is significant, they seem to make a big deal out of it. But I think there's a bigger philosophical question of, like, how accurate does a forecast actually need to be in order to be useful? You know, when we look at different fields like physics and engineering, they're using models to get spaceships into orbit. You know, they have principles that they use to determine when a model will meet the level of accuracy that they need for a specific application. We don't really have a parallel set of principles for that in biology. So I think encouraging people to think about like, is this forecast actually accurate enough for it to be used in this context is an important question we need to be asking and forecasting too yeah. 

  Marty Martin  1:25:43  

  Yeah, yeah, that's a great point as a Floridian with some experience of hurricane forecasts, I'll take some not so great ones over nothing.

  Cameron Ghalambor  1:25:54  

  Well and I think was it like Richard Lewontin, I think said: "It's a good thing they don't use population genetics to send people to the moon."

  Katie Lotterhos  1:26:03  

  Did he?  I gotta look that up. That's a good quote.

  Marty Martin  1:26:08  

  Awesome. Well, thank you so much.

  Cameron Ghalambor  1:26:09  

  Yeah, thanks for taking the time.

  Katie Lotterhos  1:26:11  

  This was really fun. Thank you so much. You you.

  Cameron Ghalambor  1:26:32  

  Thanks for listening to this episode. If you like what you hear, let us know via Bluesky, Twitter, Facebook, Instagram, LinkedIn threads or leave a review wherever you get your podcasts, and if you don't like something, we'd love to know that too, all feedback is good feedback.

  Marty Martin  1:26:47  

  Thank you to Steve Lane, who manages the website, and Molly Magid for producing the episode.

  Cameron Ghalambor  1:26:51  

  Thanks also to Caroline Merriman for help with social media, Brianna Longo for producing our cover art, and Clayton Glasgow, our Blogger. Check out his work on our Substack page.

  Marty Martin  1:27:01  

  Thank you to the College of Public Health at the University of South Florida. Our Substack and Patreon subscribers and the National Science Foundation for support

  Cameron Ghalambor  1:27:09  

  Music on the episode is from Podington Bear and Tieren Costello.