Transcript | Functional Phenomics: Studying Root Physiology Using Affordable Open Source Tools
The transcript below was generated using https://otter.ai and lightly edited. If you notice any inaccuracies please comment and add important correction within the text.
Hi, everyone, and welcome to the next webinar in our Plantae Webinar series. My name is Katie Rogers, and I am your host for today's webinar. Before we get started, I'd like to go over a few things just to make sure you get the most out of attending today's webinar. If you have questions during today's webinar, let us know using the GoToWebinar chatbox. Questions are submitted anonymously and will be read and answered periodically during the webinar as we transition from one topic to the next. If you're experiencing technical issues, please let us know about those using the chatbox or by emailing me at firstname.lastname@example.org
A recording of this webinar will be made available along with all of the associated materials within a few days. This webinar was brought to you by the American Society of Plant Biologists. I would like to give a special thank you to all of our ASPB members who are here today.
Your ASPB membership dues help support and make these webinars possible. For any of you who have not yet joined ASPB, you can join today using the discount code Webinar10 to receive a 10% discount on registration.
Today's webinar was organized by ASPB's Environmental Ecological Plant Physiology section.
Andy VanLooke is currently Chair of the EEPP section and is joining us today to take your questions and moderate the discussion. I'm going to hand it over to Andy now to share a little bit more about the EEPP section and to introduce our speaker.
Thank you, Katie, for the introduction and for getting us set up. My name is Andy VanLooke. As Katie mentioned, I'm the current chair of the Environmental Ecological Plant Physiology section at ASPB.
I'm joined today by a couple of other members of the leadership of that section, so I'd like to introduce Jen Robinson. Who is our Vice-Chair Emily Heaton as our secretary, and Ana Locke, who is our newly elected outreach officer who's helped us promote this webinar. Hopefully, you've come by the announcements via her work on Twitter and other platforms.
So we're excited to have you here today. We're excited to have this pilot launch of our webinar series coordinated by our section. And we want to hear back from many of you as we can about how we can include more at work, webinar content like this to make the society in this section more useful to our members.
Briefly about the EEPP, we're a relatively new section of ASPB. We've been around for just a couple of years, and we're piloting this webinar program as a way to start getting more of our members to talk to each other about what the work they're doing, and the exciting opportunities that are out there in terms of collaboration across the space of our community.
Our second goal is to do this integration of plant-level responses to stresses both biotic and abiotic, under field and laboratory conditions. So we can put molecular physiology, kind of one of them the main thrust of ASPB in the context of ecology. And we can provide this basic science for taking other processes from the leaf scale and the roots up to canopy ecosystem and regional skills for either managed systems. Like some of the ones that earlier I'll be talking about today or natural vegetation.
And so Larry York is joining us today to help us get started on meeting one of those particular goals. That is putting roots in an ecological context. He's got a webinar put together about his work and phenotyping roots and how that relates to ecosystem processes. He's also going to tell us about what he thinks in terms of open science opportunities and collaboration among the scientific community.
A little bit about Larry, Larry is an assistant professor at the Noble Research Institute, where he works on plant and agricultural questions with a focus on the interface between the crops in the soil.
He is really pushing the cutting edge in terms of being able to determine what's going on below ground in a way that we couldn't do just a few years ago
I'm going to hold off on any more hype about what Larry doesn't let him go ahead and give us the meat and potatoes on that. So without any further hold up here, I think we're going to let Larry tell us about his work. Go ahead, Larry.
Alright, thank you, Andy.
Let me turn on my webcam first.
Thanks for the introduction. Here I am.
Thanks for this opportunity to give the first presentation and this webinar series for the EEPP. I think that environmental and ecological plant physiology is a really important area, but sometimes it's sort of hidden by molecular physiology. So I'm happy to have this opportunity. I did want to kind of tie into sort of our Twitter campaign, and I promised that I would put some of the handles and ideas from a Twitter thread on my whiteboard. And here they are. So many of you may see your names. And I especially wanted to highlight a tweet from Mary Williams @PlantTeaching about gifting students with an ASPB membership as a great way to promote their careers.
And with that, let me turn on my presentation.
Can y'all confirm that? Do you see my talk?
Okay, and then I think I have a laser pointer here as well. So today, I'm going to be talking to you about functional phenomics. An area that I've been exploring quite a lot and studying group physiology using affordable open-source tools.
Just a little bit about where I'm at. I'm currently here in Ardmore, Oklahoma, and we are located at the National divide between the humid and wet east and relatively dry west. So it's a pretty interesting environment.
And here's a picture of the campus at the Noble Research Institute. We are a nonprofit research organization with a focus on agriculture about 400 employees. My lab is right about here. And you're all welcome to come and see us anytime you like.
So, what is functional phenomics, this idea of mix and then exploring as a way to relate phenotypes to their function, and combining big data which is relative to plant physiology. So what I do in my work is start off with considering what a root system, ideotype is. And ideotype is the ideal phenotype for a given environment that would maximize fitness or performance. And that is that we could use this to work with plant breeding. Once we consider what would be the ideotype for a given environment, say, for example, deep, deep roots are widely believed to be suitable for drought, right?
So once you know that you would want to develop a phenotyping platform. And now, with a phenotyping platform, you're essentially trying to devise a way to measure the traits of interest in as high throughput as possible and with as great of accuracy as possible. So we work on that side quite a bit.
Once you have a phenotyping platform, generally with phenomics, we're phenotyping large populations. And these can be different types, whether diversity panels, it could be across species within species mapping populations. What's important to me is to get phenotypic diversity. That's what we're after. And also, my work the functional side comes in because I don't just measure one plant property, I measure several. So, for example, I might study a lot of the root properties, but also some shoot properties like at least shoot mass, and I really don't think you can study roots or shoots about studying the other. So I highly encourage you to look at the other half if you're focused on one or the other.
Once we have this will have some type of distribution of phenotypic variation.
As you as a lot of you may know, we often hope this is a normal distribution but may or may not be. What a significant focus of plant phenotyping a lot, a lot of the assumption is that we're going to these mappings, so we're trying to associate genetic regions, whether QTL or snips, with the phenotypic state values.
So that's all exciting work. But what's more interesting to me personally is trying to correlate through study statistics linking five traits or plant phenes, a phene as an elemental unit of phenotype, and trying to link those to utility. And in my experiments This stage is largely data mining, you might say, or, I guess, the nice way to say that is hypothesis generation.
And then, to test the hypotheses that we might come up with, we move to more detailed physiological studies where we can explore these. These plant traits and more deep detail possibly live controlled environmental conditions, like draw vs. not draw some type of stress, condition, and do measurements and more detail because generally phenotyping plot platforms for high throughput screens don't allow a lot of detail. And crucially rethinking all this to simulation modeling, because computer simulations are a way to put all of our ideas about how the planet works, how the environment works, or at least the most important aspects of how they work into a testable, computational form. So that we can test our ideas, and it can also be used as a way to prioritize your research by determining which plant trades or whichever mental processes have the most impact.
And all this feeds back to our idea of what the ideotype should be. And based on that, it could all change our plot our platform. And basically, we have this cycle of creating knowledge. That's basically what I've been trying to do in my lab here at Noble for the past three years, and working with our plant breeders to actually get some of these trades into pipelines here.
Now in the context of roots, to understand which retreats are important, you have to know a little bit about soil.
So I'll briefly describe that. So in the context of soil, we have nutrients that are mobile and mobile, and mobile nutrients like you see here are those that can absorb and desorb to soil particles. And what this means is that they're relatively and mobile and that has the roots take them up. As you can see here, the depletion zone will be relatively small. So remember that immobile nutrients will have small depletion zones.
When you think about something that's more mobile and soil like nitrate, it doesn't absorb to soil particles so strongly. And so as the roots take it up, it can travel from farther away. And so you'll have larger, deep depletion zones. So remember that immobile nutrients have small locations/zones, and mobile has large depletion zones.
Now, once you know this a little bit about soil and there's a lot more to know about soil than that, trust me, but we have to consider what does a root system looks like? So I would just ask you to picture our root system in your mind, and we'll see how you do.
Some of you may have an idea of a root system, something like this. I think this is kind of how I thought about roots back then, and I guess about 2006 or so when I started to get involved with research just sort of a mass of noodles with little form.
And of course, this is true. There's a lot more structure to roof systems than this. But in fact, noodles are a pretty good analogy for roots I have this series of photographs of spaghetti noodles that I took to use to explain some concepts to you.
So any plant that grows from a seed, generally the first organ that emerges from the seed, is actually the root of the radical, or the taproot or the primary root. Those are all names for the same thing.
And this taproot will grow down, and over time lateral roots will emerge or branch roots.
And if you look at variation within a species or across species, you'll see that some plants may have few lateral roots, and some plants may have many lateral roots.
And so why does this variation matter? This is where the functional side comes in.
Well, if you imagine a mobile nutrient like phosphate like I explained before, it would have small depletion zones. So this pink area around the noodle is the depletion zone for phosphate. And the area of this depletion zone will largely determine the amount of phosphate that can be taken up.
Now for a plant with more lateral roots, this area phosphate depletion is greater. And thus, you might conclude that plants have more lateral roots that would have greater phosphate uptake.
But now let's think about a mobile nutrient like nitrate. So here, you can see the larger depletion zones of nitrate with the plant a few lateral roots. And if you model this in the same way with the plant with many lateral roots, you can see that basically, the total area of nitrate depletion has not increased. As well, even though you have more roots, they're not necessarily more effective.
And the important thing here is to remember that roots like noodles; they have a cost. And so when you're spending more to get the same amount of nitrate, the efficiency, as viewed from the standpoint of the plant or the root system is less. So we spend a lot of time thinking about this concept of cost benefits.
There are some other root traits that are important. So, for example, if you imagine these roots that have emerged here, maybe they're basal roots, like an avene. If you imagine something like water or nitrate that can become a more deep resource throughout the season than having plants with a steeper angled root, it may be beneficial to grow deeper can access that that resource.
On the other hand, if you imagine roots that have grown more shallow, and you think about some of the fertilizers, we apply, like phosphorus. Like I said before, phosphate is relatively and mobile in the soil. And not only that, even natural systems phosphate will accumulate in the topsoil due to deep mining by the roots and decay of the chute above ground, leading to accumulation of phosphate over time. So some nutrients are expected to be more abundant in the topsoil. And in that case, having shallow roots may be beneficial.
So that was just sort of a brief primer of some root traits and how they would matter in the context of different soil resources that differ and their mobility.
And then with our work with root phenotyping, it is to take these ideas and get measurements of these trades and try to understand how they are working and filled conditions.
So I started to take pictures of roots to try to measure traits. Many, many years ago, I guess about 10 years ago, and actually my idea at the time, I didn't actually know about image analysis. I just thought that if I took a picture now I could measure the trade by hand later on the computer screen and save myself time in the field. But later, I did learn about image analysis. And over time, my methods for taking images of roots have evolved.
A couple of years ago, while during my postdoc in Nottingham, England, I thought that the picture you see on your screen was pretty good with a dark background, white tag for identity, a circle for the scale.
But what I found out is that in practice, computer software has a difficult time trying to identify what is the root and what's the background and images like these. So when I got to noble I took a kind of a gamble to try a few new things and one of those was using these type of machine vision cameras that you can see here, they don't have any buttons or anything nothing for a consumer to point and shoot with. But they are made to interface with the computer so very convenient. This one is a monochrome camera. And I've experimented with using a backlight as the background. So these are actually just led flat panels to use for ceiling lights.
And we've devised ways to eclipse and replace the system to put root crowns that are excavated from the fill the so-called shovel mix method or root crown phenotyping. And we have an imager software that the interfaces with this brand of camera of Bosler camera, the users can change some of the camera settings, the same location file names, but one of the more innovative uses that really helped us with our methods is that we like to use barcodes for all of our samples, which I encourage you all to do to help you keep track and to age or downstream analysis.
And what we can do is that we have a barcode scanner. And when we scan the barcode for a sample, it automatically triggers the image acquisition and saves the file with the barcode identity as the file name. So then when we do our image analysis, we already have the identity and we can quickly go to the statistics directly. And this preprint is available now. I'm hoping that it'll be accepted for publication soon. And the software is available on Zenodo.org which I'll talk a little bit more about later.
And this all was developed by on and by Seethepalli and in my group here at Noble.
Now once we have all these sorts of black and white images of roots, we need to analyze them. So we also have this analyzer software that is basically just made for batch analysis of the pictures and you have a few settings you can set the thresholding level and the diameter classes.
And basically, it goes through a directory of images and gets the data. One nice feature is that we also help to these feature images so that you can use them in your presentations to kind of show what traits are being measured. And this can process hundreds of images in just a few minutes.
Also available as a note. So essentially, we have the raw image from a camera. This is how the image analysis works. And when we process it, we first we threshold the image which you can barely tell a difference because we've optimized the image damaging so much that the raw image almost already is a segmented image.
Once we have a segment of images, we can calculate things like the convex hole which gives us some idea of the exploration extent of a root crown. We calculate the distance map, which is the distance of every black pixel from a white pixel. And from that, you can see these ridges form. And by climbing those ridges, we can identify the root skeleton. And from the skeletons, you can get the root length, which is a very important metric.
And also from that distance macro getting the diameters and trades like that. And we're using these disconnected components or holes. If you zoom in, you can see these basically are forming where there's a background that surrounded by roots. And we think these are a useful index for how branchy the root system is.
So we get this suite of trades and the software is evolving pretty quickly. And this is the current version is a graphical user interface version. And I think I'm going to risk doing a quick demonstration of that. Let me see if I can.
Okay, yeah, I think I'm now and our imager software and we can toggle through the different views of the image. So these are scan roots so this new version will also be able to accommodate scan through images and you can draw regions of interest and you can choose the thresholding value, some diameter ranges, you can set the DPI and choose what's shown in delta image. And when we do the analysis, you see that we get some information the histogram about the diameters and we can also drag around and look at it zoom in and I'll toggle with space
From the wrong image, the segmented image and finally the features with the skeleton overlaid.
So that's where we're kind of going with that. And we're hoping to have that ready over the next year.
So hope that will be a useful tool for the community. So back to the presentation.
Based on this type of technology, we realized that we could also do some shoot phenotyping here at Noble. So another example of affordable phenotyping, by the way, the last platform I showed the rise division crown for root crown phenotyping cost about $1,500 to make and I've, I've released all the plans for that.
So for the shoot imaging, we're using a color camera and a black background, a light, and we're getting these images of wheat seedlings are have an experiment we have growth chambers, either with or without heat stress. And so we're getting pictures over time, again, because we've optimized the lighting conditions. So it's pretty easy to segment the images and get data. And our plan is to also make it easy to use software for doing this analysis.
Now, one interesting use case for the root phenotyping platform was this work with trying to look at the effects of cotton root rot, which is a disease that affects many species. And here we are looking at alfalfa.
So here's an alfalfa field that's infested with cotton root rot. What you can see happen is that you have these disease rings where the disease is spreading and so you have a disease front. In the background, you have asymptomatic plants that don't show the disease.
You have the survivors that had had the disease, but survived and what we did is we excavated root grounds from the survivors in asymptomatic plants. And what we were able to show with this platform is that we were able to distinguish the disease status using the multivariate analysis, linear discriminant analysis of all the traits. And basically what we found is that the survivors were able to compensate for the loss of the taproot here, you can see a taproot on the left, the survivor doesn't have this and it compensates by having more and longer fine lateral roots.
And I wanted to use that that last week I showed you lift the alfalfa cotton rock, as as an example of how we do open science.
So open science is a philosophy to make research accessible to everyone.
And there are some components to open science like open access, open data, and open-source and on their own. They're great, but open science is when you combine all these together.
So what is open access basically means to make journal articles available to everyone. It also generally means that they're free to use as long as you give credit. One way to do this is with preprints. and preprints are where you post your manuscript online when you're submitting to journals or before for open access when you publish in that journal, most journals require an article processing charge.
And many new journals are open access only. But a caveat here is to watch out for predatory journals. predatory journals are open access journals that are popping up for mostly for profits. And there's been a lot of negative effects of that. So if you're unsure about a journal, I would encourage you to ask people in your community about it.
So with this paper about copyright law, you can see that we posted the pre preprint and after it was published on the bio archive was where we posted the preprint and it was published in five items journal and you can see that by archive was able to automatically link it.
So one fear with pre presses you to know what happens when it publishes and in fact, they're trying to make the DOIs is the digital object identifiers be linked. And you can see here the published version in Phytobiomes journal.
And this is a great way to communicate your research.
Open Data means making your raw data and my opinion your statistical code available to everyone. So data includes things like numeric tabular data, images, data from instruments, and this is useful for the validation of new methods for other people who are developing allows for reproducibility and validation of results.
And really importantly, by doing this and having this level of transparency, you really build trust among scientists and also between scientists and our stakeholders, including the public.
One way that we've done this is, for example, I mentioned this zonodo.org before, they're related to the CERN particle accelerator in Europe, and they freely allow you to upload data set. So here, we uploaded all the images use for that cotton robot, study all the data and all the R code to do the analysis to this site. And you can see that people are actually downloading it and I hope that they're checking it and if I made an error that will tell me but it's all there for anybody to see.
Open Source refers to making software code and also, in my opinion, hardware plans available. So open source refers primarily to program code, but us. For me, it's also useful to make your hardware plans available so that people can really reproduce what you said you did. And of course, I'm making programs that are executables all ready to run, it's also useful.
And this can include web applications. So an example of this, GitHub is a good way to share software code. So this is for the riser vision analyzer that does the image analysis. So all the code is available, then we actually made the executable available on sonoda.org. Another good thing about Zonodo, relative to GitHub is that it tracks the downloads. So you have some metrics to show if your tools are being used.
And finally, as I said, I also tried to release all of our hardware plans, even down to having a parts list available.
So I encourage all of you to try and open science on approach. It does take a little more work.
But in the hand, I think that really helps make your work more visible and more useful to the community.
So I have a few minutes remaining, and I want to talk about some other plot platforms that are more physiology-based. This is pretty new work in my lab that we're very excited about.
And one of those is to look at root respiration.
So if you imagine, yeah, roots are living tissue, so they're respiring just like we do accelerating co2, and how much co2 they require is related to their activity. And so you have roots that might require a lot of roots that could require less is a little carbon dioxide, more molecules, and the idea is that there may be some luxury respiration maybe for currently unknown purposes that the plant could do without So, that, in other words, is that if we can reduce respiration without having any negative effects on other root functions, then this may be a promising way to reduce the metabolic burden of the root system.
And I kind of see that as a corollary, a lot of people are talking about increasing photosynthesis. Well, another way to look at that is to do is to decrease throughout the respiration.
So, we developed a platform where we can grow a lot of plants and hydroponics which I realized is kind of cheating for roots, but this is sort of our first step and this is the work of Dr. Jai Childwall in my group.
So we take these plants out of the hydroponic setup, and we have three like or infrared gas analyzers that we cut off the root system and place it in these chambers. attach them to these gas pathways to be read into the analyzer, and use that we use a bead bath to maintain the temperature.
And so a team of eight people, which was kind of a lot for us to get together, we're able to do 600 plants in a day. So it was hard work, but it showed that it was possible. And we've now collected four replicates of data for we diversity panel with about 300 members. It's relatively high throughput, high throughput for reward, honestly.
And so there's of course, there's a relationship of total group length with total CO2 flux. But what we're really interested in is the residuals. So the plants that are able to have more roots but less little flux, and other words that specific group respiration so we divide the total flux by the root link or root mass
This is a graph of the heritability for several trades that we're getting out of this platform. For example, here a shoe to dry away with a heritability around point six and specific respiration.
By the length and by mass have good heritability.
So we're excited about this trade. And we're proceeding with genetic mapping and starting to try to study you the possible utility and more depth.
And other physiological trade that we're measuring is ion uptake or nutrient uptake. So we know that transporters at the molecular level are responsible for uptake. And most of our measurements of uptake are done at the segment or system levels.
And so we're interested in trying to understand really how uptake scales across these different levels and how different abilities uptake nutrients at this level would affect the performance of crops in the field.
So we did some meta-analysis to look at the uptake rate in the literature. And there's not a lot of work on it because I think a) roots are hard to study and then b) physiological traits of roots are even harder to study. But what we found is that there is work and that if you look along with the species shown that it looks like there's some variation. For example, the Imax is a maximum up uptake rate of nutrients by roots. There's some variation for this across species and the literature.
And so, we had a project to try to develop at least a method to measure ion uptake of, in this case, multiple nutrients. And this is the work of Dr. Marcus Griffiths in my group. So we have now a medium-throughput platform. To be honest, I don't want to call it high, high throughput, but we can do measurements for about 50 plants a day. And so if you stagger your applications, you could do entire populations theoretically. So we have 224, channel peristaltic pumps, and we have a platform where we grow the plants in hydroponics and transfer them to this measurement plot platform where one pump adds the known concentration starting solution, and then we're sampling with another pump overtime to get the nutrient solution and over time, the roots are taking up nutrients and so we're measuring the nutrient solution which will be declining and concentration so we're measuring the depletion in solution, but most reasonable people accept that the depletion in the solution is equivalent to the net uptake of nutrients by the roots.
So what we found out is that if you look at these boxplots across all the maize NAM population parents. So this is a kind of diverse panel of maize maze lines. There's quite a bit of variation for in this case and nitrate depletion. And when we calculate the heritability, it's over 0.5. And so trying to think about like, why would there be differences and uptake ability?
So we know that you have a transporter, it's able to take up a nutrient at some rate, but maybe there's a different type of transporter.
And a different level of the same gene. We know there are different genes for transporters, but what we don't know is if there are different variants with energy, different levels that have different properties.
There's only one example in the literature of this, so maybe this one's twice as fast as another way is to have more abundance of transporters and that should have greater uptake. But what we don't know again is how exactly do you see the scale from fewer to more transports and how will that affect the total uptake.
And finally, there are other processes that are happening at the same time for example, for nitrate, you have to be pumping protons from the outside of the cell, I mean from the inside of the cell to the outside to drive the process and then you have things like assimilation. So in the long run, like we tried to, what we would like to try to understand is exactly which processes are driving differences and, and uptake.
And finally, I'll conclude with this concept of the whole plant, economic spectrum. So this started off with the leaf economic spectrum in 2004. And basically, the idea is that you have plants that make plants you have plants that make with a fast style that makes leaves that are more new, nutritious with greater photosynthetic rates, but with shorter lifespans the plants with long leaves.
Longer lifespans, but less nitrogen, less photosynthesis. And people extended this idea to roots as well. There's an economic spectrum. But the idea has not been used in crops a lot. So in this work with maize, we've tried to look at this concept of a plant, economic spectrum. And what we uncovered is that it does seem like football for both leaves and roots, that there's some type of economic spectrum that matches what people have seen in the literature, which is typically measured across natural species like tree species.
So we've uncovered the spectra within across species. And we've shown for example, with the principal component analysis that the first component is related to the leaf economic spectrum. And then the second component is related to the root economic spectrum. And you can see that here with these correlation patterns among leverage and then separately among retreats, they sort of naturally clustered in this way. So this is another area that we'd like to follow up on. I think it's a good tie in with ecology and physiology because it's also used a natural system we read research quite a bit.
So yeah, with this slide, I'd like to kind of come conclude that I've talked about the functional phenomics pipeline, from thinking about video types to developing and phenotyping platforms phenotyping populations and looking at variation and correlations. I didn't talk a lot about these two. But I think that these are also really important aspects. And so I hope that these ideas might be useful to some of you out there.
And I'd like to have a couple of slides after this to put together some some some advertisement for a conference and such, but I'd like to acknowledge the people in my lab, they've been a real pleasure to work with my first three years leading a group I've been lucky to have good people to work with.
And my funding sources primarily Noble Research Institute, the state of Oklahoma and the USDA have all funded aspects of this work.
And I would like to let you know about our Lloyd Noble Scholar and Plant Science Program. Applications are due by December 31. And this is a summer a paid summer internship for I believe, junior and senior undergraduates. So just Google this or go to noble.org. If you'd like to know more about that.
This is to let you know about the Rooting 2020 conference in Nottingham, so it's a nice symposium focusing on roots, and that's Rooting2020.com. Registration opens in January.
And finally, the Phenome 2020 meeting is coming up in February and registration is is is opened up. And I've been to this meeting for the first event. This is the fourth year and I've been each time it's always a very, very well done and interesting meeting.
And I'll leave with some of the Twitter handles for myself and my group and the EEPP group from ASPB.
That's all I have and I think that Andy is going to moderate some discussions. I hope we can have some nice discussion about fun moments and open science. Thank you.
Great, thank you very much, Larry. And I have audio so you can hear my clapping. Perhaps there are other people clapping from around the world along with this as well. So I really appreciate the presentation I learned so much, Larry, but I'm going to ask the questions of other folks first, and I think we might be able to fill up most of the time. Oh, I see we have clapping coming into the question bar too. Good.
Great job. Okay, so I'll go through the questions here as well as I can and we'll see how many we can get through. The first question we have is from Amir. Amir wants you to clarify whether or not you're washing the soil off of the roots before imaging. And then when you imaging with the RhizoVision system, how do you actually get the image? I mean, which environment you image the root, so you're doing it in the air on the soil, and then the RSA will be in the same orientation as in the soil. Thanks. So please clarify here from here about how you get the soil afterward.
Yeah, those are good questions that I should have clarified in the presentation. So these are excavated root crowns, so they're dug up with a shovel, we use method washing off with water. And, of course, when you're digging the root crown up, it's definitely not the whole root system, you're breaking it off from the rest of the roots. So we're losing roots there. We also know we're losing roots home, we're washing off the soil. There are some other methods to release the soil. I'm aware of a recent publication using air spades to take off the soil. I think that any of these methods you'll lose the fight the find roots, and there's no way around that really this method is designed to be high throughput. So it's the highest throughput film based root method that I'm aware of. And the important thing is that just despite and most orientation, of course, you can see my hands but with gravity, you know, you'll be pulling some of the roots down
Some roots are more or less rigid. And I actually have a test that you could call it the York test or rigidity if you wanted to. But take the root count, turn it upside down, the roots that don't change their shape are the ones that are more or less rigid, right. And if you do this, you'll see that there are some roots, especially like the thicker, noble roots amaze, or, for example, hold their shape quite well. In other words, don't. So we're losing some information, but it's fast. But the important thing is that these traits that we measure, I didn't show this data today, but for even a week that has fine a fine root system, we're getting pretty strong correlations with things like shoot mass and shoot nitrogen content from these methods.
So in short, you are losing roots. They do change the orientation, but we still think that the traits are relevant.
Great, thank you, Larry. I think the next question we have is a little more on the technical side.
How do we measure root branching density? And when you measure the root angle do you measure the angles of all the roots branching from the main root?
So with our analyzer tool, you do get branch points from the topology and root tips, and the holes, which we all think our index for root branching, but you may not be measuring group branching itself, and that's a common thing and plant phenotyping that a lot of things we're measuring are more of a proxy than the actual tray. And for our purposes, where we're looking for relative differences among genotypes, maybe that's good enough. If you want greater accuracy, you might have to do more detailed work. Some software, like for plant seedlings, such as root math, will actually get down you can actually label the roots as which class of root if it's a lateral group, and then you'd be able to get the angle of that specific lateral root which are our software does not give that type of detailed root by root metric.
Okay, thanks, Larry. Another question, what if we use noninvasive imaging like an x-ray?
Well, I think I've worked with that, right and Nottingham, using smaller vessels sizes, I think it's a very, very promising tech technology. And I'm, I've been kind of keeping it on my radar. The challenges are things like segmenting the roots from the soil volume is still pretty, pretty difficult. And especially for finer roots. It's very difficult. So the processing times are slow, a lot of manual intervention is typically done.
So from my standpoint, it's also very expensive, like a typical extra unit costs at least $500,000.
And so I think it's promising, but it will take time to go mainstream but very promising.
Okay, great. We have another couple of methods questions I'm going to combine into one here, Larry,
Are you able to work with smaller roots of plants like Arabidopsis growing in smaller media? And can you monitor group growth over time?
Yeah, so I have done that I use the sort of miniature platform as what I showed for the root crowns to measure medicago. I haven't done Arabidopsis this but medicago sort of small too. And we are able to use that same type of camera system to get through images every 10 minutes and look at root growth over time. So that's, that's one way to do it. There are some other plot platforms available as well. So if that person who asked a question and wants to send me an email, I can try to follow up.
Okay, good that there was a question.
want to take a step back from the technical side and ask you a little bit about your career path. So can you just tell us about how you got to where you are now and the biggest thing you learned kind of moving through the various stages from undergrad to grad school and things that aren't kind of technical about your current role?
Sure. I guess I guess the most important thing is never to give up. Actually and an undergraduate at one point I flunked out of undergraduate more because I stopped going so apparently if you stop going to school without withdrawing, you'll get all E'sor F's. So but I did have a second chance to go back and finish up. And I guess for me, and I don't think there's a one size fits all for everybody, but for me, I generally just took open opportunities, you know, when I saw its chance, I usually just said yes, you know, and, and took it. And that has and then to be proactive, you know about finding those opportunities, you know, be ready and accepting a lot of opportunities and be proactive about finding them as well. So for example, I was, as a graduate student, talking to the person who ended up being my first postdoc supervisor. Keeping doors open, you know is, is important, but guess the most important thing is to keep trying.
I think those are important traits of someone who studies roots too.
So, Peggy wants to know about the variation of the ecology of this system. So, how much variation is there root situ to respiration among different crop plants, different crop plants under different stresses? Like how does drought or temperature affect it?
So I think that respiration is a very responsive trait. I think it's remarkable in and just in this weak diversity panel that we studied there's I can't remember it's at least three, threefold variation, and that's just within a species. And so and ecology, you know, that's kind of one of my beefs with ecology, and I have a PhD in ecology is that most of it are done at the species level. And then our, the premise of all our work is that there's tons of variation within species. So I would suspect that among species, there's more variation than what I saw on this crop diversity panel.
And in my own work so far, respiration I've seen with low and high nitrogen, that there's a response of respiration, and that respiration across these diversity panels of maize is correlated to the uptake of nitrogen. So it is a responsive important trait and that's kind of why I'm proposing a sort of a, you know, sort of a master trade that controls may be many other trades.
And then what about the environmental variation and temperature and drought and moisture conditions?
So I know it's responsive to temperature, which is why we use the BB to control the temperature. I don't know the extent of that. Like, how much can you change it just based on temperature? And I'm not personally familiar with the effects of drought, but I assume it would be really responsive to that as well.
So I have a question coming back to something detailed on RhizoVision. So is the raw image used in rising of vision to a composite of various angles, or do you choose it from only one promising angle from which to take the pictures?
Yeah, that's a good question. In my work, I've always chosen subjectively. So based on the user's opinion, widest view, so if you kind of twist that around by I will kind of pick out what's what, what's, what's the widest view. And that seems to work. We do have some work, it's also available as a preprint with Felix Fritchie, and Alena Zare, looking at multiple camera angles, so it had five cameras from five different angles. And that works so that using five angles helped in the and out in the analysis like it has extra information, but it was fairly marginal.
So I, in my own work, I think that for the moment, a single angle is sufficient. But I know that Alex Books and Georgia is working on a 3D, I guess, sort of a 3D method for root crown phenotyping. So I'll be interested to see, you know, sort of a comparison of what the 3d data gives you compared to an isolated single perspective.
I have a question coming from Jane. And she's asking about open data in and sharing with other African universities. Right? So she says. So this is quite new and African universities, including her own the University of Eldoret, it's where she is now. How do we partner and sharing skills and knowledge and the usage of these open-source tools? I am very interested in phenomics.
Well, I guess I mean, one reason why I wanted to highlight this in my talk is that I feel like this isn't this, this level of open science that I'm trying to do, I don't really think is done very, very much. And there's no reason really cannot do it. I don't think there are many bars I mean, to sharing your data and our code, I imagine that most people would have access to Zonodo.org. So if you want to share your data and our code there's no bar to it.
And beyond that, getting access to two-part two partners, I guess as a matter of just reaching out to people.
But beyond that, yeah, this is the type of community the discussion that I would like to. Great. I'd like to think so far the webinar might be helping get that started to. Okay, we're running a little bit low on time here. We have five minutes left, and maybe we'll ask, unfortunately, one more question. So Larry, I just want to give you a second to kind of think about your final thoughts. And the follow up here. And I'll get you one more question in that window. Trying to find out the hope the one that we can get the most interest in a lot of good questions. Okay.
Well, we'll tell you will tell you kind of a different one. What is the main difference between RhizoVision and other software tools for imaging like DIRT and GIAR roots?
Thanks just from a mere, Okay, yeah, I'm familiar with both those software and in fact, I was a co-author on the original DIRT paper.
So honestly, there's not a ton of differences like for GR Root is more general-purpose there and our analyzer plot platform are built for root crowns. DIRT focuses currently on a cloud-based platform. And we wanted a tool to you know, that users could use on their own computer. And then also we have extended hours to be more useful for scanned images are as a sort of a dual purpose for both root crowns and also for scan root damages like you might get soil core roots.
So we're trying to basically optimize for our own purposes within the lab, a lot of the use cases of different types of workflows that you would have within the software that may not be conducive to cloud-based and not specific to root crowns. So SARS is basically sort of a little more general purpose. And we have the ability to keep up to date, like, for example, geo roots are no longer as far as I understand, being kept up to date. So we're hoping to maintain, our software over time as well.
Great. So, Larry, we have I believe, three minutes left, I wanted to give you a chance to sort of summarize or your thoughts or add anything from the discussion that kind of came to your mind and also want to prompt you as well. You alluded to the Twitter contest on your board, but I'm not sure we know who the winner is. So if you wanted to mention that as well and our last few minutes to go by, wave my hand or somebody, actually, what I ended up not actually I'm gonna tweet a picture out of that of this board right after we finish here and ended up putting basically all the handles up. Okay.
So it's an open-source when I was disappointed that you didn't want to draw the picture of your own face that I tweeted at the last minute, but I'll take it as a loss. This is my face. Okay, you got it.
Okay, Larry. So here in the last few minutes and anything else you want to leave it to the webinar community and any other plugs, I think I should give one to please consider joining plant biology. I see some folks who said they'd like to do that and the EEPP section does not require membership in an ASPB you can join our section for just $5, the link that Katie showed earlier, and we'll send it out, and the follow-up. So if you'd like to hear more information like this, we'd appreciate your membership do that. And Larry to the last two minutes of yours.
Okay, I guess I would just from a lot of the questions, I can see there's a lot of interest in tools for roots. And so someone had to point out that I think there's really no magic bullet for roots and that basically any way that you choose to study them is based on your questions.
There are some, some some some websites available that correlate different software tools that you can look through to choose ones that might be best for you. And I can't I can put that on Twitter as well. I think it's now quantitative plants used to plant image analysis. And then, as far as open science goes, I really encourage anyone still listening to push their boundaries and try to do as much as I can. Share their data, there are only benefits as far as I can tell, to being more open.
Okay, great. Well, I'll wrap up my portion, again, thank you very much, Larry, there's there are quite a few questions here that we didn't end up getting to. And also quite a few thank yous down below. So I'm going to encourage those who did send questions that to go ahead and send you a follow-up email, you may get quite a few here in a short burst and hopefully, you can follow through with the ones that you're able to answer.
And I appreciate you taking the time to do this today and spending I think a really good example of what we're trying to do with EEPP. So we hope to do more of this in the future. Please, for those who are still hanging on the line if you have ideas about other topics that are similar to this or you know, colleagues or you know yourself and you can do something as nice and useful as what Larry has done here. Please let us know.
Thank you. Thanks, everyone.
Thanks, everyone. All of these, this video at all.
These discussion points will also be posted on the plant a so so be on the lookout for the recording and continuing discussion.
Thank you all.
Transcribed by https://otter.ai