Deep Roots & Big Data: The Evolution of Our Crops: A conversation with Dr. Peter Morrell, Professor of Agronomy and Plant Genetics
Good day and welcome to the University of Minnesota podcast Minnesota CropCast. I'm your host Dave Nicholine, University of Minnesota Extension Educator in Field Crops. I'm here today with my cohost Doctor. Seth Nave, University of Minnesota Extension soybean specialist. And joining us today in the studio, Seth is one of our University of Minnesota Agronomy Department faculty, Peter Morrell.
Dave:And Peter, I don't want to make sure that I don't mistake in terms of your proper academic title, but I believe it's a professor in computational biology.
Peter:Yep, that's, you got it, exactly.
Dave:Well, we're gonna have you talk a little bit more and explain what computational biology, but let's back up a little bit. So folks out here in the University of Minnesota and in the surrounding states and the world for that matter, may not know who you are a little bit. So can you give us a little bit of background information about yourself, where you grew up, where you went to school, and how in the world you ended up here at the University of Minnesota? So we'll chronologically start at the beginning if that's okay.
Peter:Sure. Yeah, I so I grew up in Southern Oklahoma in a relatively small town called Tishomingo, which which means war chief in Chickasaw or Choctaw. So the town is named after a Chickasaw warrior, and the town was originally the capital of the Chickasaw Nation. And I grew up in town, very close to downtown, but my grandfather and my mother's family had a ranch about 12 miles south of there. So my first involvement with agriculture and ranching was through my grandparents.
Peter:I gather this town wasn't a very large town or isn't a very large town per se. It's about 3,500 people. It's become relatively well known because, Blake Shelton, who's a popular country musician, has set up businesses and things there, and I think his former wife, Miranda Lambert and Gwen Stefani and bunch of there there are lots of entertainment folks that go there now, but it's still a town of about 3,500 people.
Seth:It's no Branson, Missouri yet.
Peter:No. It's there's one or two bars downtown that have live music now though.
Dave:So this experience in the ranch, how did that translate into biology and the agriculture or was it just one of those subjects that you studied in school and said, hey, I like this, I want to do this.
Peter:Well, the nice thing about this area was it was a great place to be outdoors. So there were lots of creeks and rivers and streams and lots of places to play. So, and it's warm in Southern Oklahoma, so we spent all of our time outside, and that was my first interest in biology. Later on, I was an exchange student to Australia when I was about 16, and my biology professor there told us you know, if you're interested in going into biology, plant science is really the place to go, because we need a lot more plants than animals in terms of food, and in terms of human food. And so plant biology offers a lot of research opportunities.
Peter:And it was really her influence that sent me into plant biology.
Seth:Where did you go to school?
Peter:I so there is a there is a local two year school I started to call it a community college, but it's not. A two year school in my hometown that was set up to train Chickasaw and Choctaw men. They now are co ed, and they've grown a lot. I went there for two years because I came back from Australia, I hadn't enrolled in school yet. And then I went to the University of Oklahoma, which is in Norman, which is the it's not the ag school.
Peter:The ag school is Oklahoma State, which is in Stillwater. I was actually born in Norman. My dad was a graduate student there at the time. So I was sorta, to a certain extent going home when I went to the University of Oklahoma at Norman. But yeah, that's where I did my undergrad.
Dave:And graduate school after that?
Peter:I went to went to a graduate program at the Claremont Colleges, which are private schools in Claremont, California. So the best known of those schools is Pomona College, but there's also a there's graduate also a program there, and I worked with a scientist named Lauren Reisberg, who was a really, really good evolutionary biologist and got me off to a really good start in terms of research and evolutionary genetics primarily.
Dave:Okay, and did that end up with a master's or a doctorate level or did you go beyond that?
Peter:I got a PhD there and then and then as a postdoc, I worked with a couple of people who moved around. So I was at Texas A and M for a short while, and then at the University of Georgia, and then I came back to California to UC Riverside, and then the person I worked with there moved to UC Irvine, so I was in a lot of different places over a course of a couple years.
Seth:So pull a thread through this. Just give us the what connected all this, or what do you did you just walk from one opportunity into the next, or is there something that kind of you feel like in looking backwards kind of pulled you all the way through this whole system to get you where you are today?
Peter:Well, the research I had done as a graduate student was primarily on natural plant populations, but I wanted to move more in the ag realm just because I thought there were lots of interesting applications of genetic research in ag. I mean, obvious applications. And so ultimately, when I ended up at UC Riverside, I worked with a professor there, Michael Clegg, who had done a lot of early work on barley and wild barley in terms of DNA sequence level analysis, and I really wanted to go into sort of more direct applications related to, you know, at least somewhat related to agriculture, and that was what really got me interested in sort of the current approaches that my lab still uses today.
Seth:So you feel like you're somewhat on the basic side, but you really feel like you have, you're driven by these larger questions around agriculture and food production and things like that, sounds like.
Peter:Yeah, that's been an interest for a long time.
Dave:Well, you spoke about a lot of what I would
Peter:call the
Dave:southern locations. What got you up here to the land of ice and snow to the University of Minnesota? We're quite far north of obviously of Oklahoma. Was there opportunities in a lab or was there money available or just how did you end up here at the U?
Peter:Well, I had I had left a postdoc and gone to work for Monsanto for a couple of years, but while I was there Oh, St. Louis then. I was working out of out of St. Louis, yeah. Okay.
Peter:But while I was at Monsanto, University of Minnesota was recruiting for a couple of different positions, and I knew people here because I had done barley research. And most of the most of the barley genetics in North America, it happens on the St. Paul campus, and so I knew Brian Stefansson and Gary Muehlbauer and Kevin Smith and those guys from from prior research.
Seth:Okay. So I think maybe I'm gonna ask a little bit about Monsanto because and maybe you have NDAs and you can't talk about any of this, but
Peter:It's been a while.
Seth:Top secret stuff. But, you know, they're obviously a for profit company that's, you know, that was interested in producing products to sell to farmers. So what what kind of work did you do there and how how did you position yourself in a company like Monsanto? What what kinds of things did you offer somebody a corporate like them?
Peter:Well, the thing that I was hired to do, position that I was hired for was was to carry out forward genetics research to try to find genes primarily in corn, in maize, that could be used to transform maize and create basically transgenic products that would increase yield and productivity.
Seth:Okay, so help, we've got to go back to the basics here. So define forward genetics for our So
Peter:with forward genetics, you're essentially trying to use traditional genetics approaches like crosses and segregating progeny from a set of parents so that you can identify locations in the genome that might contribute to individual traits. So I think when Bob Stupar was here, he talked about QTL mapping. Yep. Yes. That's one aspect of forward genetics.
Seth:Okay, perfect. Perfect. We just want to get everybody on the same level with this. So now you're at the University of Minnesota and you were hired, what was your job description originally when applied here for this, your current position? And how do you feel like you're fulfilling those original requests from the department?
Seth:Do you think you fit well with what that original job description indicated?
Peter:I think the goal at the time was to bring in someone who had more experience with sort of computational analysis of biological data. And I would sort of add that all biology is essentially computational now. No one uses an abacus to add up their field results. So it's all computed. But what my group has primarily worked on is various types of analysis at the DNA sequence level.
Peter:So you could call it genomics. Some of it is probably better identified as population genetics, so understanding population level diversity in various crops and how that affects their utility. But it's not that computational biology is fundamentally different than biology in general.
Seth:And so you're really you're interested in a very broad way about genes and their functions in individual plants and populations of plants and how that can contribute to varieties and types, other pools of plants and animals. Is that fair?
Peter:Yeah, it is. I mean, I'll give an example. One of the things that we've worked on for a number of years, yield increasing crop yield is generally good for everyone. Lower inputs and higher output. And so basically our thesis has been that if you could make healthier plants, you could increase yield without increasing inputs, without increasing environmental impact, and everybody gets more productivity for the same input.
Peter:But one way to do that is to think about the genetic composition of the plants we actually work with. So this is gonna get slightly technical for just a minute, but, you know, all new mutations have some effect on an organism and they vary from being lethal, you know, so killing the organism to having slightly harmful effects to being positive. And there's a distribution of those. And what you think that distribution looks like depends on sort of whether you're glass half full, glass half empty, I guess. But there are a lot of harmful mutations that segregate within populations.
Peter:And one of the things that we've worked on is ways to identify those variants just from DNA sequence and eliminate them. So increasing yield by removing things that are negative, and so removing genetic variants that are harmful to the organism is a way that we can fundamentally increase productivity.
Seth:So let me just put this in a classical or please put this in a classical context of getting rid of these deleterious mutations versus just allowing the process to occur by selection or by natural selection. So how, I assume it sounds like you're just trying to speed up the process, right?
Peter:We're trying to speed up the process. So just to give you some scale, so every human genome, for example, has somewhere between one and three harmful mutations that would be lethal if they were homozygous. But you have several thousand mutations that are slightly harmful, And plants are pretty similar, a few thousand mutations that are slightly harmful. So if we can use DNA sequence to identify those slightly harmful mutations, we can choose progeny or offspring from a cross, a biparental cross, that have fewer of those harmful mutations than either of their parents did.
Seth:And is there any commonality across these? Can you just go in could we just go in and zap a thousand of these things all at once and have them have them just be gone with all the bad stuff?
Peter:That's been suggested, but but it's not very practical. There are tools to go in and zap individual mutations, but it's it's sort of one at a time. You know, a lot of people have heard of CRISPR for example and there are base editors, but yeah. So let's just back up a
Dave:little bit and talk about crops specifically Here in Minnesota, what field crops have you been spending a little bit more time with? Where's your emphasis been? And then the second part of my question, what is the end product that is coming out of your lab? Is it a technique, a knowledge, and you can say this is how to do this and you turn it over to the plant breeders and the plant breeders and say, oh, well, I will follow this recommendation or this technique and so forth because they're actually putting the wheels on the car per se. Maybe there's a two part question there.
Dave:So as I suggested earlier, I have done a lot
Peter:of research on barley and we have ongoing projects on barley. The other things that we've been funded to work on recently are soybean. We've had multiple projects on soybean and pea or especially pea grown for protein. We've had recent projects on. And then we've had funding from ultimately pass through funding from the Gates Foundation to fund research on cowpea, which is an important crop in especially in Western Africa.
Dave:Okay. But again, now you're that that end product that I was talking about before that, you know, because if you're not putting the wheels on the bus, are you saying this is how the wheels should be assembled? This is what we know and so forth. So, it's it's that that knowledge that they can incorporate within their own program?
Peter:Yeah, so we've written both we've we've written them both of course sort of methodology as to how we think this could be done. We've also demonstrated that it can be done, you know, by by looking at genomic prediction populations in barley and determining that indeed when you select for yield, you end up with fewer deleterious mutations. And we published that several years ago now, about five years ago. But we also have written software that allows people to implement these approaches. And right now, we're starting to work with language model based approaches to enhance these.
Peter:So we were pretty good at finding harmful mutations in coding sequence, in the actual gene space within a genome. Plants have a lot of space outside the genes, which are hard to annotate. And so we've been working on trying to identify generalizable ways to annotate variants outside of coding space so that we find a lot more of what's potentially harmful.
Seth:So I assume that you're working in the most efficient sphere you can. You're looking at plants that give you the most potential gain for your input, you know, that you can provide, or your efforts that go into it. But more broadly, are so diverse and we have so much different kinds of genetics and how those ploidies and out crossing and selfing and all the different variations in how plants manage themselves over generations. So how do you see your work? How can you optimize your what's your biggest gain for your inputs in these very different kinds of systems?
Peter:Well, it's a good question. So we have argued that you get the biggest gain out of these approaches that I've talked about, like removing harmful mutations, when you're working with an inbred crop. Because in soybean, in barley, you don't sell a hybrid version of that crop. In hybrid system, you're often one of the major theories that explains heterosis is that you're offsetting these harmful mutations. So you still will get more genetic gain if the inbreds are healthier.
Peter:That's been shown over the history of maize breeding that healthier inbreds lead to more vigorous hybrids still. But even when you eliminate a lot of the harmful things that are in the inbreds. But I think when in soybean, in barley, in wheat, where where you're primarily selling an inbred line, the more you remove those things that are causing a reduction in fitness for the plant, the more yield potential you get out of it with little sort of mitigating factors, you know, associated with it. You might lose a little bit of heterosis if you eliminate enough of the harmful mutations, for example, right?
Seth:So I think of I think about soybean as really highly inbred, but also bottlenecked. We don't have a lot of genetic genetic variation, and we've had plant breeders for fifty years that have been crossing and re crossing and re crossing and re crossing these plants. Is there any kind of any of these deleterious mutations that have just of nodded themselves into these genomes, do you think, that are just really is there any of these that get tied to important genes or that are linked in some way or that maybe structurally can't be altered? I'm wondering about what the big breakthrough you might make if you can find some of these things and then zap them like I said.
Peter:Well you mentioned zapping earlier and and yeah the so there are two types of harmful mutations that you could imagine and in a species like soybean where the diversity is quite low so so the average pair of soybean lines is differs at about every thousand space pair. They're which is about the same diversity as humans, by the way. But but it's not very much, and so there there a lot of the variation I'm talking about is harmful variation, for example, in barley, we're seeing as what we call segregating variation. It's still present in the population. In species like soybean, you could have fixed deleterious mutations, and those would be candidates for zapping, because once you have fixed a harmful mutation, it decreases the fitness of that line, and there's no going back, essentially.
Peter:The only way to revert it would be to have an individual that was segregating for the alternate variant. So there is some new evidence, I think published a day or two ago, in a major paper about how inbred lines can sometimes overcome some of these things. I think it was a species called Amazon molly, which is a fish that can overcome some of these effects of inbreeding. But yeah, once you fix those differences, they're really they're they're there to stay, and they do increase the fitness, and they and they can be really problematic.
Seth:So let's think about impacts and ROI. This just occurred to me this morning. I was in a meeting, there's a lot of interest in the university doing a better job, And there's more questions about what the return on investment for state funds and grant funds and things like that. So since you work a little bit on the more basic side, I assume that a lot of what you do just takes a little bit more time to get to reach the end, you know, the person on the street in Minnesota, the farmer, whoever that is out there. So maybe try to explain to us a little bit how you might how you think about how you impact, you know, and especially stakeholders here in Minnesota.
Peter:Well, talked about the harmful mutation piece, so I'll mention something that's related, but not exactly the same. We have an ongoing research project on the nature of recombination. So when you make it when a breeder makes a cross between two inbred lines, they're hoping to combine the best traits of those two parents into into offspring that are higher yielding than either parent was. But the ability to do that is determined entirely by where by where recombination events occur in the progeny that come from that cross. So we've been working on an approach that asks about the landscape of recombination, specifically within biparenteral crosses, and thinking about essentially the breeding value of an individual line that contributes to how it changes the recombinational landscape.
Peter:So maybe it produces more recombination in a region that's near something, you know, on a two genes that you'd like to disassociate, or maybe it actually reduces recombination in an area where you'd like to preserve a set of co adapted traits. So we're working right now on an approach that allows us to take biparenteral crosses from breeding data, and give a very direct readout of where desirable recombination is occurring and also potentially the genetic factors that are controlling those recombination events.
Dave:So these predictive tools actually from a plant breeder standpoint help them speed things up a little bit versus the tried and true methodical thing like I can only observe in the field.
Peter:Yeah, I mean, it's, we're essentially trying to take data that almost everybody collects anyway. Like, like Aaron Lorenz, when he, when he makes a cross in soybeans, at some point, he genotypes all those progeny. But, and nowadays there are more sophisticated tools being used, but traditionally that data went into a spreadsheet, and somebody looked at the spreadsheet and pointed at a spot where the colors changed and the alleles they found and said, Here's a desirable recombinant. But we're trying to automate those sort of processes so that it's much easier to find those desirable recombination events, for example.
Seth:Let me make one more jump here and then we'll probably let you go. But I'm, you know, everybody, every day, well, every hour, every minute, people are talking about AI. It sounds like what you've been doing is really AI before we talked about AI in some ways, right? Is this partially explain some of the things that you do in your lab, then automate some of the things that we've been doing manually, and include new data sources that maybe weren't included in decisions, and you know, then the machine spits out an output for you. Is it is it fair to say that that you're what you're doing is is is basically an AI approach?
Peter:They're certainly related. I mean, there's sort of lineal descent between the idea of automating approaches, you know, and making tools that work very generally to solve problems and then being able to apply them to lots of different organisms. So we try to do things that can be applied across a variety of organisms. We we are using AI a lot more for that, but like, you know, a year ago, if we wanted to write a Python program to to do some of this analysis I was describing, you had to write the program by hand and it took, you know, sometimes it took all afternoon to get something prototype that worked. You can spit something out of a AI check client that will get the process started.
Peter:It's usually not perfect in a half an hour, but those things there's sort of a continuum between those things. The the step changes are not quite as steep when you see them from up close, think.
Seth:Okay, so I think we're going to wrap up here, but let's think about the future a little bit. What excites you in terms of what you're doing in your lab or your approach or the outputs or,
Peter:you
Seth:know, what what is kind of driving you and what's kind of on the on the horizon that keeps you coming back for another day here in the office?
Peter:Well, haven't mentioned training students, but my lab has been very involved in training both undergraduate, graduate students, and postdocs. And I have to say that the skill set that they learned in our lab doing the sort of computational biology work I've described, it is very general. A lot of the students leave for jobs in logistics or we have a former undergrad who works at NASA as a programmer, one who works for Apple. You know, the skill set is very broadly applicable, but it's really exciting to see what students do with the skills they learn. And most of the grad students that come out of my lab and several of the postdocs now work at PepsiCo and Syngenta especially, those two companies.
Peter:But but they're they really find a broad range of applications for the skill set they learn.
Seth:And do you think that that those companies hiring both the ag companies and and those other companies appreciate the the obviously the type of education, but even the problem sets that they were working on, that they were real life issues that they were working on rather than super theoretical, even even those that ends up end up in NASA or places like that. I assume that they they look beyond just their their computing skills to other other traits too. So do you think that those things are are were helpful for them?
Peter:I think so. I mean, the students tell me that the papers they read to understand the science we were doing, and then the things they wrote as students really helped them think very clearly about what they're doing. But they also the publications that students produce as undergrads or grad students really demonstrate their skill set. Once they're inside the company, that may not be as easy to do. So their ability to have some demonstrated work product is very important for them, more than I ever sort of realized when I started this.
Peter:It it makes a difference. I I have a guy who's worked on geographic information systems, and he's made maps for us. And he tells me, when I'm working with a new client, I send them a couple of papers and stuff because it's in the public domain, it's on the web. I can say, look, this is a product I created. And so it is it really helps students to to have demonstrated skill sets.
Seth:That's a great example of another value of these publications. I think some of the public don't necessarily understand that the value and getting those work pieces out into the domain, the variable values that are attributed to those. So I think I really appreciate that.
Dave:Yeah, would call it the multiplier effect, you know, in terms of that, just the tools and the technology from a classical plant brain situation, but all these other things that, excuse me, can affect industry and opportunities and a lot of things. Well, as Seth mentioned, we are at the close here today. So Peter, we'd like to thank you for stopping by today and visiting with a little bit more. I have a better understanding now of computational biology. Even if it's from a guy way down south in Oklahoma.
Dave:I think you translate it pretty well here to the northern climate with all of us Scandinavians and German and so forth up here. Think the science is a science so we appreciate that. So thanks again for stopping by.
Peter:Thank
Dave:you Any last words?
Peter:Just thank you for the invitation. I enjoyed talking to you today and really appreciate it.
Dave:Okay, thank you very much. Well this has been another edition of the University of Minnesota CropCast. I've been your host Dave Nicolai along with my co host Doctor. Seth Nave, University of Minnesota extension specialist in soybeans and we'll look forward to visiting with you next time. Thank you.
